{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "00c3b786",
   "metadata": {},
   "source": [
    "# Cvičenie 3 - Regularizácia a generalizácia modelov"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fab6f758",
   "metadata": {},
   "source": [
    "## Príprava prostredia\n",
    "V aktivovanom virtuálnom prostredí spustite nasledujúci príkaz:\n",
    "\n",
    "Zdrojový kód\n",
    "pip install numpy pandas matplotlib \"scikit-learn>=1.4\"\n",
    "\n",
    "O čom je cvičenie\n",
    "Nízka chyba na trénovacích dátach sama osebe nezaručuje, že model bude dobre predikovať aj na nových dátach. Model sa totiž môže prispôsobiť nielen skutočným vzťahom v dátach, ale aj šumu a náhodným osobitostiam trénovacej vzorky. Regularizácia obmedzuje efektívnu zložitosť modelu a môže zlepšiť jeho schopnosť zovšeobecňovať na nové dáta, spravidla za cenu horšieho prispôsobenia trénovacím dátam. Predstavuje tak jeden z hlavných nástrojov na riadenie kompromisu medzi vychýlením a rozptylom (bias-variance trade-off). Príliš slabá regularizácia môže ponechať model pretrénovaný, zatiaľ čo príliš silná môže viesť k podtrénovaniu.\n",
    "\n",
    "Cvičenie sa zameriava na rozpoznanie pretrénovania a na jeho obmedzenie pomocou regularizácie. Porovnáte regularizáciu s L1 penalizáciou, L2 penalizáciou a ich kombináciou, preskúmate ich vplyv na koeficienty modelu a vhodnú silu regularizácie zvolíte pomocou krížovej validácie. Porovnaním trénovacej a validačnej chyby pri rôznych hodnotách sily regularizácie sa naučíte rozpoznávať prejavy pretrénovania a podtrénovania a lepšie porozumiete kompromisu medzi vychýlením a rozptylom.\n",
    "\n",
    "Ciele\n",
    "Rozpoznať prejavy pretrénovania a podtrénovania modelov na základe porovnania trénovacej a validačnej chyby.\n",
    "Vysvetliť rozdiely medzi L1 penalizáciou, L2 penalizáciou a ich kombináciou a opísať ich vplyv na koeficienty modelu.\n",
    "Pomocou krížovej validácie zvoliť vhodné hyperparametre regularizácie a vyhodnotiť validačnú výkonnosť modelu.\n",
    "Interpretovať regularizačnú cestu a vysvetliť vplyv jednotlivých penalizácií na koeficienty korelovaných príznakov.\n",
    "Posúdiť generalizačnú schopnosť zvoleného modelu na nezávislých testovacích dátach a porovnať výsledky s odhadom získaným krížovou validáciou.\n",
    "Dátová sada\n",
    "Na tomto cvičení budeme pracovať s dátovou sadou Diabetes, ktorá obsahuje 442 pozorovaní a 10 numerických príznakov, napríklad vek, index telesnej hmotnosti, krvný tlak a vybrané hodnoty z krvných testov. Cieľovou premennou je číselný ukazovateľ progresie ochorenia po jednom roku, preto ide o regresnú úlohu.\n",
    "\n",
    "Vzhľadom na pomerne malý počet pozorovaní je táto dátová sada v spojení s dostatočne zložitým modelom vhodná na názornú ukážku pretrénovania. Pomocou polynomického rozšírenia vytvoríme mocniny a súčiny pôvodných príznakov, čím zvýšime počet vstupných príznakov a flexibilitu modelu. Model sa tak môže výraznejšie prispôsobiť aj šumu a náhodným osobitostiam trénovacích dát. Následne budeme sledovať, ako regularizácia obmedzuje jeho efektívnu zložitosť a za akých podmienok zlepšuje schopnosť zovšeobecňovať na nové dáta.\n",
    "\n",
    "Funkcia load_diabetes vracia príznaky už centrované a škálované na celej dátovej sade (predvolené scaled=True). Pri vlastných dátach sa škálovanie učí iba z trénovacích dát, preto je StandardScaler súčasťou každej Pipeline aj v tomto cvičení.\n",
    "\n",
    "Dátová sada slúži ako didaktický príklad na demonštráciu princípov strojového učenia. Vytvorený model preto nie je určený na klinické rozhodovanie ani na predikciu zdravotného stavu konkrétnych pacientov.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c280dcd3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training set: (353, 10)\n",
      "Test set: (89, 10)\n",
      "      mean    std    min    max\n",
      "age  0.001  0.046 -0.107  0.111\n",
      "sex  0.000  0.048 -0.045  0.051\n",
      "bmi  0.002  0.047 -0.089  0.161\n",
      "bp   0.001  0.048 -0.112  0.132\n",
      "s1  -0.001  0.048 -0.109  0.154\n",
      "s2  -0.001  0.048 -0.116  0.199\n",
      "s3  -0.001  0.047 -0.102  0.181\n",
      "s4   0.000  0.048 -0.076  0.185\n",
      "s5   0.001  0.048 -0.126  0.134\n",
      "s6   0.002  0.048 -0.138  0.136\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.datasets import load_diabetes\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "RANDOM_STATE = 42\n",
    "\n",
    "# load the Diabetes dataset\n",
    "data = load_diabetes(as_frame=True)\n",
    "X, y = data.data, data.target\n",
    "\n",
    "# split data into train and test sets\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=RANDOM_STATE)\n",
    "\n",
    "print(f\"Training set: {X_train.shape}\")\n",
    "print(f\"Test set: {X_test.shape}\")\n",
    "print(X_train.describe().T[[\"mean\", \"std\", \"min\", \"max\"]].round(3))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a6c17e3b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "9c79fd7c",
   "metadata": {},
   "source": [
    "## Teoretické minimum\n",
    "V tomto cvičení používame regularizáciu, ktorá do účelovej funkcie pridáva k trénovacej chybe penalizáciu veľkosti koeficientov modelu. Pri učení preto hľadáme kompromis medzi prispôsobením trénovacím dátam a obmedzením veľkosti koeficientov. Cieľom je zlepšiť predikcie na nových dátach - zlepšenie však nie je zaručené a overujeme ho pomocou validácie.\n",
    "\n",
    "Lasso (L1 regularizácia) - penalizuje súčet absolútnych hodnôt koeficientov. Niektoré koeficienty môže nastaviť presne na nulu, čím vytvára riedky model a vykonáva výber príznakov. Pri silne korelovaných vstupoch môže výber závisieť od konkrétnej trénovacej vzorky. Nulový koeficient preto nie je dôkazom vecnej nedôležitosti premennej. Pri polynomickom rozšírení sa výber týka jednotlivých členov - pôvodných príznakov, ich mocnín a interakcií - nie automaticky celých pôvodných premenných.\n",
    "\n",
    "Ridge (L2 regularizácia) - penalizuje súčet štvorcov koeficientov. Zmenšuje ich celkovú L2 normu, spravidla ich však nenastavuje presne na nulu, takže nevykonáva výber príznakov. Pomáha stabilizovať odhady najmä pri silne korelovaných vstupoch a má tendenciu rozdeliť ich príspevok medzi viaceré koeficienty. Neznamená to, že každý jednotlivý koeficient musí s rastúcou hodnotou hyperparametra alpha monotónne klesať v absolútnej hodnote.\n",
    "\n",
    "Elastic Net (kombinácia L1 a L2) - spája možnosť nulovania koeficientov s ich stabilizáciou. Pri silne korelovaných príznakoch má väčšiu tendenciu ponechať skupinu súvisiacich príznakov spoločne než samotné Lasso. V implementácii scikit-learn hyperparameter alpha určuje celkovú silu penalizácie a hyperparameter l1_ratio pomer jej L1 a L2 zložky. Hodnoty bližšie k 1 zdôrazňujú L1 penalizáciu, hodnoty bližšie k 0 zasa L2 penalizáciu. Pri l1_ratio=1 dostávame Lasso, pri l1_ratio=0 zostáva iba L2 penalizácia. Na samotnú L2 regularizáciu v cvičení používame Ridge.\n",
    "\n",
    "Výber sily regularizácie - väčšie alpha znamená silnejšiu penalizáciu v rámci danej metódy a pri nezmenených ostatných nastaveniach. Príliš slabá regularizácia nemusí potlačiť pretrénovanie, naopak príliš silná regularizácia môže viesť k podtrénovaniu. Vhodnú hodnotu vyberáme podľa validačnej chyby pomocou krížovej validácie. Pri Elastic Net môžeme rovnakým spôsobom vyberať aj hyperparameter l1_ratio. Rovnaké číselné alpha pri rôznych metódach neznamená rovnako silnú penalizáciu, pretože sa líši tvar penalizácie aj škálovanie jednotlivých členov účelovej funkcie.\n",
    "\n",
    "Poznámka\n",
    "Penalizácia pôsobí priamo na veľkosť koeficientov. Ak majú vstupy rozdielne jednotky, rovnaká zmena predikcie môže vyžadovať veľmi odlišné koeficienty. Bez škálovania teda penalizácia závisí aj od voľby jednotiek. StandardScaler preto patrí do tej istej Pipeline ako model.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "353cc026",
   "metadata": {},
   "source": [
    "## Postup\n",
    "Krok 1: Demonštrácia a diagnostika pretrénovania\n",
    "Samotná vysoká chyba natrénovaného modelu ešte neznamená pretrénovanie. Typickým prejavom pretrénovania je výrazná generalizačná medzera - model dosahuje nízku chybu na trénovacích dátach, ale podstatne vyššiu chybu na validačných dátach.\n",
    "\n",
    "Zložitosť modelu budeme zvyšovať polynomickým rozšírením príznakov. S rastúcim stupňom pribúdajú mocniny a interakcie pôvodných príznakov, čím rastie počet koeficientov lineárnej regresie a zároveň aj samotná flexibilita modelu. Pre každý stupeň porovnáme trénovaciu a validačnú chybu neregularizovaného modelu a budeme sledovať, ako sa mení generalizačná medzera."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e03d08c",
   "metadata": {},
   "source": [
    "### Úloha 1.1 - Vplyv zložitosti modelu na pretrénovanie\n",
    "Preskúmajte, ako stupeň polynómu ovplyvňuje trénovaciu a validačnú chybu. Pre každú hodnotu hyperparametra degree zo zoznamu [1, 2, 3] skonštruujte Pipeline s krokmi PolynomialFeatures, StandardScaler a LinearRegression. Pri polynomickom rozšírení nastavte include_bias=False.\n",
    "\n",
    "Modely vyhodnoťte pomocou päťnásobnej krížovej validácie s rovnakými foldmi, a to iba na trénovacej množine, konkrétne X_train, y_train. Testovaciu množinu ponechajte na záverečné vyhodnotenie zvoleného modelu. Použite cross_validate(..., scoring=\"neg_root_mean_squared_error\", return_train_score=True). Vrátené skóre má opačné znamienko než štandardná hodnota metriky RMSE, preto pred interpretáciou výsledkov zmeňte jeho znamienko. Pre každý stupeň vypočítajte priemer trénovacích a validačných hodnôt metriky RMSE cez päť foldov.\n",
    "\n",
    "Doplňte časti kódu označené komentárom TODO v kostre nižšie.\n",
    "\n",
    "Na základe výsledkov posúďte, či zvýšenie stupňa polynómu zlepšuje generalizačnú schopnosť modelu, alebo vykazuje prejavy pretrénovania. Sledujte najmä situáciu, keď trénovacia chyba klesá, ale validačná chyba rastie.\n",
    "\n",
    "Upozornenie\n",
    "Najnižšia trénovacia chyba nezaručuje najlepšiu generalizáciu. Zložitejší model sa môže prispôsobiť aj šumu v trénovacích dátach a na nových pozorovaniach dosahovať horšie výsledky. Model preto vyberajte podľa validačnej chyby, trénovaciu chybu používajte spolu s validačnou na diagnostiku pretrénovania."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "d768c02f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   degree  n_features  train_RMSE  validation_RMSE  generalization_gap\n",
      "0       1          10       53.35            55.39                2.05\n",
      "1       2          65       47.63            62.50               14.87\n",
      "2       3         285       14.10          1402.18             1388.08\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.model_selection import KFold, cross_validate\n",
    "from sklearn.pipeline import Pipeline\n",
    "from sklearn.preprocessing import PolynomialFeatures, StandardScaler\n",
    "\n",
    "cv = KFold(n_splits=5, shuffle=True, random_state=RANDOM_STATE)\n",
    "rows = []\n",
    "\n",
    "for degree in [1, 2, 3]:\n",
    "    # TODO 1: create a pipeline with the steps \"poly\", \"scaler\" and \"reg\"\n",
    "    model = Pipeline(\n",
    "        [\n",
    "            (\"poly\", PolynomialFeatures(degree=degree, include_bias=False)),\n",
    "            (\"scaler\", StandardScaler()),\n",
    "            (\"reg\", LinearRegression()),\n",
    "        ]\n",
    "    )\n",
    "\n",
    "    # TODO 2: use cross_validate to compute the RMSE score for the training and validation sets\n",
    "    scores = (\n",
    "        cross_validate(\n",
    "            model,\n",
    "            X_train,\n",
    "            y_train,\n",
    "            cv=cv,\n",
    "            scoring=\"neg_root_mean_squared_error\",\n",
    "            return_train_score=True,\n",
    "        )\n",
    "    )\n",
    "\n",
    "    # convert negative scores to RMSE\n",
    "    train_rmse = -scores[\"train_score\"].mean()\n",
    "    validation_rmse = -scores[\"test_score\"].mean()\n",
    "    # count the generated polynomial features\n",
    "    n_features = model.named_steps[\"poly\"].fit_transform(X_train).shape[1]\n",
    "\n",
    "    rows.append(\n",
    "        {\n",
    "            \"degree\": degree,\n",
    "            \"n_features\": n_features,\n",
    "            \"train_RMSE\": train_rmse,\n",
    "            \"validation_RMSE\": validation_rmse,\n",
    "        }\n",
    "    )\n",
    "\n",
    "overfit_results = pd.DataFrame(rows)\n",
    "\n",
    "overfit_results[\"generalization_gap\"] = (\n",
    "    overfit_results[\"validation_RMSE\"] - overfit_results[\"train_RMSE\"]\n",
    ")\n",
    "print(overfit_results.round(2))\n",
    "\n",
    "# plot RMSE score by polynomial degree\n",
    "fig, ax = plt.subplots(figsize=(8, 5))\n",
    "ax.plot(\n",
    "    overfit_results[\"degree\"],\n",
    "    overfit_results[\"train_RMSE\"],\n",
    "    marker=\"o\",\n",
    "    label=\"Training RMSE\",\n",
    ")\n",
    "ax.plot(\n",
    "    overfit_results[\"degree\"],\n",
    "    overfit_results[\"validation_RMSE\"],\n",
    "    marker=\"s\",\n",
    "    label=\"Validation RMSE\",\n",
    ")\n",
    "\n",
    "# log scale makes large differences in RMSE easier to compare\n",
    "ax.set_yscale(\"log\")\n",
    "\n",
    "ax.set_xticks(overfit_results[\"degree\"])\n",
    "ax.set_xlabel(\"Polynomial degree\")\n",
    "ax.set_ylabel(\"RMSE (log scale)\")\n",
    "ax.set_title(\"Training and validation error by model complexity\")\n",
    "ax.legend()\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b9c0083f",
   "metadata": {},
   "source": [
    "## Krok 2: Porovnanie regularizačných metód Ridge, Lasso a Elastic Net\n",
    "V predchádzajúcom kroku sme pozorovali, že zvýšenie zložitosti modelu môže znížiť trénovaciu chybu, ale zároveň zhoršiť generalizáciu. Pri učení lineárnej regresie model každému príznaku priradí koeficient, teda váhu určujúcu jeho príspevok k predikcii. V našom prípade majú vlastné váhy aj vytvorené mocniny a interakcie príznakov. Teraz preskúmame, ako regularizácia obmedzuje veľkosť týchto váh a môže tým zmierniť pretrénovanie.\n",
    "\n",
    "Ridge regresia (L2) zmenšuje veľkosť koeficientov bez toho, aby ich spravidla vynulovala. Pri korelovaných príznakoch má tendenciu rozdeľovať ich príspevok k predikcii medzi viaceré koeficienty. Lasso regresia (L1) môže niektoré koeficienty nastaviť presne na nulu, čím vykonáva implicitný výber príznakov. Pri silne korelovaných vstupoch však môže malá zmena dát viesť k výberu iného príznaku. Elastic Net kombinuje L1 a L2 penalizáciu, čo umožňuje vytvoriť riedky model, teda model s časťou nulových koeficientov, a zároveň môže zvýšiť stabilitu pri korelovaných vstupoch.\n",
    "\n",
    "Pri rovnakom polynomickom rozšírení porovnáme, ako tieto metódy ovplyvňujú validačnú chybu, veľkosť koeficientov a počet nenulových členov modelu."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "af1d6d2e",
   "metadata": {},
   "source": [
    "### Úloha 2.1 - Porovnanie vplyvu regularizácie na chybu modelu\n",
    "Preskúmajte, ako regularizácia ovplyvňuje trénovaciu a validačnú chybu modelu. Porovnajte lineárnu regresiu bez regularizácie (OLS) s metódami Ridge, Lasso a Elastic Net. Vo všetkých prípadoch nastavte v kroku PolynomialFeatures hyperparameter degree=2 a použite rovnaký postup polynomického rozšírenia a štandardizácie v rámci Pipeline. Krížovú validáciu vykonajte iba na trénovacej množine, konkrétne X_train, y_train, s rovnakým rozdelením do piatich foldov ako v Kroku 1.\n",
    "\n",
    "Použite tieto pevne zvolené hodnoty hyperparametrov: Ridge(alpha=10.0), Lasso(alpha=1.0) a ElasticNet(alpha=1.0, l1_ratio=0.5). Hyperparameter alpha určuje silu regularizácie. Pri metóde Elastic Net hyperparameter l1_ratio navyše určuje pomer L1 a L2 zložky penalizácie.\n",
    "\n",
    "Pre jednotlivé modely vypočítajte a porovnajte priemernú trénovaciu a validačnú hodnotu metriky RMSE cez päť foldov. Variabilitu validačnej chyby vyjadrite smerodajnou odchýlkou validačných hodnôt RMSE medzi foldmi. Posúďte, ako sa oproti neregularizovanému modelu zmenila validačná chyba a rozdiel medzi trénovacou a validačnou chybou. Vplyv regularizácie na koeficienty preskúmate v Úlohe 2.2.\n",
    "\n",
    "Výsledky interpretujte ako demonštráciu účinku regularizácie pri konkrétnych nastaveniach. Hodnoty hyperparametrov zatiaľ neboli ladené, preto z porovnania nemožno vyvodiť všeobecné poradie kvality metód. Rovnaká číselná hodnota hyperparametra alpha navyše nepredstavuje pri rôznych metódach ekvivalentnú penalizáciu.\n",
    "\n",
    "Doplňte časti kódu označené komentárom TODO v kostre nižšie.\n",
    "\n",
    "Nižšia priemerná validačná RMSE sama osebe nepreukazuje lepšiu generalizačnú schopnosť modelu. Pri interpretácii zohľadnite aj variabilitu výsledkov medzi foldmi, vyjadrenú smerodajnou odchýlkou. Tá však nie je intervalom spoľahlivosti ani kritériom štatistickej významnosti. Keďže modely vyhodnocujeme na rovnakých rozdeleniach dát, podrobnejšie porovnanie má vychádzať z párových rozdielov hodnôt metriky RMSE na jednotlivých foldoch. Závery sa vzťahujú na použité konfigurácie, nie na všeobecnú prevahu jednej regularizačnej metódy.\n",
    "\n",
    "Dokumentácia: Ridge, Lasso, ElasticNet, cross_validate"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "0720db0b",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                               model  train_RMSE  \\\n",
      "2  <class 'sklearn.linear_model._coordinate_desce...       50.39   \n",
      "3  <class 'sklearn.linear_model._coordinate_desce...       52.61   \n",
      "1        <class 'sklearn.linear_model._ridge.Ridge'>       49.03   \n",
      "0                                                OLS       47.63   \n",
      "\n",
      "   validation_RMSE  validation_SD  \n",
      "2            57.13           1.79  \n",
      "3            57.14           2.00  \n",
      "1            58.60           3.02  \n",
      "0            62.50           4.93  \n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import ElasticNet, Lasso, Ridge\n",
    "\n",
    "# fixed (untuned) hyperparameters to demonstrate the effect of regularization\n",
    "models = {\n",
    "    \"OLS\": LinearRegression(),\n",
    "    # TODO 1: add \"Ridge\", \"Lasso\" and \"Elastic Net\" with the hyperparameters from the assignment;\n",
    "    #         set max_iter=50_000 for Lasso and Elastic Net so that coordinate descent converges\n",
    "    Ridge: Ridge(alpha=10.0),\n",
    "    Lasso: Lasso(alpha=1.0, max_iter=50_000),\n",
    "    ElasticNet: ElasticNet(alpha=1.0, l1_ratio=0.5, max_iter=50_000),\n",
    "}\n",
    "\n",
    "rows = []\n",
    "fitted_models = {}\n",
    "\n",
    "# apply the same feature expansion and scaling to each model\n",
    "for name, regressor in models.items():\n",
    "    # TODO 2: pipeline with the steps \"poly\" (degree=2, include_bias=False), \"scaler\" and \"reg\" (= regressor)\n",
    "    pipe = Pipeline(\n",
    "        [\n",
    "            (\"poly\", PolynomialFeatures(degree=2, include_bias=False)),\n",
    "            (\"scaler\", StandardScaler()),\n",
    "            (\"reg\", regressor),\n",
    "        ]\n",
    "    )\n",
    "\n",
    "    # TODO 3: 5-fold CV on the same folds cv as in Task 1.1 (RMSE, including the training scores)\n",
    "    scores = cross_validate(\n",
    "        pipe,\n",
    "        X_train,\n",
    "        y_train,\n",
    "        cv=cv,\n",
    "        scoring=\"neg_root_mean_squared_error\",\n",
    "        return_train_score=True,\n",
    "    )\n",
    "\n",
    "    # fit on the whole training set - Task 2.2 inspects the coefficients\n",
    "    pipe.fit(X_train, y_train)\n",
    "    fitted_models[name] = pipe\n",
    "\n",
    "    # convert negative scores to RMSE and summarize fold results\n",
    "    rows.append(\n",
    "        {\n",
    "            \"model\": name,\n",
    "            \"train_RMSE\": -scores[\"train_score\"].mean(),\n",
    "            \"validation_RMSE\": -scores[\"test_score\"].mean(),\n",
    "            \"validation_SD\": (-scores[\"test_score\"]).std(ddof=1),\n",
    "        }\n",
    "    )\n",
    "\n",
    "# sort by validation RMSE\n",
    "comparison = pd.DataFrame(rows).sort_values(\"validation_RMSE\")\n",
    "print(comparison.round(2))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "088f689f",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "8f961ada",
   "metadata": {},
   "source": [
    "### Úloha 2.2 - Vplyv regularizácie na koeficienty modelu\n",
    "Preskúmajte, ako jednotlivé regularizačné metódy ovplyvňujú koeficienty modelu v porovnaní s neregularizovanou lineárnou regresiou. Z modelov uložených v premennej fitted_models získajte koeficienty a priraďte k nim názvy zodpovedajúcich polynomických príznakov. Všetky modely musia byť natrénované na rovnakej trénovacej množine X_train, y_train s rovnakým polynomickým rozšírením a rovnakou štandardizáciou.\n",
    "\n",
    "Pre jednotlivé modely porovnajte:\n",
    "\n",
    "L2 normu vektora koeficientov,\n",
    "najväčšiu absolútnu hodnotu koeficientu,\n",
    "počet koeficientov považovaných za numericky nulové podľa podmienky abs(coef) <= 1e-10.\n",
    "Do týchto výpočtov nezahŕňajte absolútny člen (intercept). Koeficienty interpretujte vzhľadom na štandardizované polynomické príznaky. Niektoré polynomické členy sú redundantné (napr. druhá mocnina dvojhodnotového príznaku sex), preto jednotlivé koeficienty neregularizovaného modelu interpretujte opatrne.\n",
    "\n",
    "Kód zároveň pre každý model vykreslí 15 koeficientov s najväčšou absolútnou hodnotou a farebne odlíši kladné a záporné hodnoty. Doplňte časti kódu označené komentárom TODO v kostre nižšie.\n",
    "\n",
    "Pomocou vypočítaných charakteristík a grafov porovnajte, ako jednotlivé metódy ovplyvnili veľkosť koeficientov. Pri metódach Lasso a Elastic Net posúďte aj to, či pri zvolených nastaveniach vznikol riedky model. Nulovanie vyhodnoťte podľa počtu numericky nulových koeficientov, pretože graf najväčších koeficientov ho nemusí zachytiť."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "85a9fa17",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3oAAAJOCAYAAADhz3V3AAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjExLjIsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvgI3uAAAAAAlwSFlzAAAPYQAAD2EBqD+naQAAjBdJREFUeJzt3QeYE9X7//2bXpTeO0gVBEUFREQFxAKIFAs2LIgFCyoqX0AEBQU72LCjKIqIgthAsGCjSVOQIkoTBOmLdNg81+f8/pMnG7K7yW52k82+X9c1sJlMZs7MpMw955z75PH5fD4DAAAAACSMvLEuAAAAAAAgugj0AAAAACDBEOgBAAAAQIIh0AMAAACABEOgBwAAAAAJhkAPAAAAABIMgR4AAAAAJBgCPQAAAABIMAR6AAAAAJBgCPQAAAAAIMEQ6CHTtmzZYpMmTbLt27cnxNH8+++/3f7s3r07ofd78eLFNnXq1HTnIXNWrlzp3ieHDx+O+LXZdT4i2U5m9ie7Jep+xTu+R2IjOTnZFixYYJMnT7YZM2b45+/Zs8dmzpxpH330kf3222+WlJTk3uvr16+PaP1ffvmlzZ8/33IifQ8sWrTIcnt5cvI5jERu2c+w+IBM+vLLL316K/3www9xdSz//fdf34cffuhLSkqK6HXvv/++259FixblyP0O1+233+4rUaJEuvOQOSNGjHDvk507d0b82uw6H5FsJzP7k1lLly51n2lvmjJlim/OnDm+AwcO5Jj9+vXXX913R6wcPnzY980337jj9/fff4dcZt26dSmOc+C0bdu2dLeRaN8jn376qW/+/Pm+eLZ3717f6aef7qtWrZqva9euvnvuucfNX7Bgga9UqVK+U0891de9e3ff22+/7fvtt9/ce/2dd96JaBs1atTwXX311b6ceLz1frzlllt88SJW5cnsOcwJn4XsfK/mBPnDCweBnGfZsmV22WWX2fLly61BgwaxLk6O0LRpU3f3F7lLTjnv77//vj366KN2/vnnW7FixWzfvn2uBuPo0aP20ksv2eWXXx73+/X222/blClTbPXq1dm6XdXe6NipJuH444932//www/t0ksvPWbZ77//3q699lo777zzrESJEimeO+mkk6xMmTKWm9x000124YUX2ltvvWXxauzYsbZw4UL7559/rHz58v75w4YNs9q1a6eo3diwYYN1797datSoEdE2OnTokC2/pTnheOdWnJuch0APKRw6dMj9UET6A4DE0KtXLzchd8lp53306NH+C87//vvPBSTXXHONNWvWzGrVqpXj9kvfuQpcFYBlFTUxP+2001ywp2Z8V155ZbqvefLJJ+2UU07JsjIhujc2S5UqlSLI8+Y3b948xbxq1aq5ppuR0s0UADkLgR7UfNd++ukne/fdd90d3htuuMGeeuqpFEdGd8x//fVX13+tevXq7q5uvnz5Uj166t/m9RHIkyePFS5c2OrWrWv16tUL2a9gyZIltmnTJqtUqZI1bNjQLR/pMuHK6Lr++OMP9zotryktOl5qf69tlCtXzk4//XQrUKBAxMdHx3vOnDnWvn17K168uLsrq3kdO3Z052Pr1q3uLqsudmfPnu3OyRlnnGFFixbNcN8a3fnv3Lmzf562Gel2VIvyyy+/uOXr1KljJ554YornV6xYYUuXLvXvv/ZNF5Q6VoECt7137153LLR827Ztw9qfwNfrmM+dO9dtq0WLFm49ovnz5s2zggUL2plnnpniPAVfjOv86/Ny8sknW+XKlUMu9/vvv9tff/3l7qIH73ekxyma0jqWoc57RvYneFn1e1OfoEsuueSY4xqtfVdwdOutt7rvrenTp7u/PTllvz7//HPr27evW58CVtVY5s+f/s+z9z2i7w69J9Oi2k1NsRKtz/yRI0fc51j9zBS4KrD57rvv3HPnnnvuMdtV7ZW2q3Vo+eDtib5TdU61TKNGjaxChQr+34qPP/7YDh48aOvWrfMHR6rRbNOmTbr7rO9/vU/03aJ9VSAWTGVbs2aNC/R1o+K4444Lua7U9kNl+/TTT91z+t0JDuB0LDdu3BgysFMAqN/zcMusfk9ly5Z15Qy3fB7VJCvA1HtQy+r3S8dZy3rfxeEe79TOV7h0vPWZ1nd48GdCx7J06dLWqlWrY16n1gMqV9euXf1lDvbtt9/6++3rM6zjoH1M6xojrfJEus/hvp88es+oT6eu6YJra7UvefPmtXPOOSeiz0I4n7mMHvOMHN9I9jNYpPsSl2LddhSx8/vvv/sGDRrkq1mzpmuvX7VqVd99993nW7VqVYrl1J9E7Z3LlSvna9++ve+UU07xnXzyyb4lS5ak2ldtzZo1rj+AN5177rm+woUL+1q1auXbsWOHfzn1Fahdu7bbdocOHXzNmzf31apVy/fRRx9FtEwo3377rSvX8uXLI1pXqD5606ZN8xUvXtx3ySWX+P77779U99tbtnr16r7KlSv7LrjgArd+Pf75558jPj5eWbQtPXfOOef4Klas6Pvnn39c+/MqVar4vv/+e1+jRo18559/vq9ChQq+SpUqub5M0eqjF8l2kpOTfY888oivSJEivsaNG7v3S8mSJX3nnXdeir49kyZN8u97ly5dXP+R/Pnz++69996Q254xY4avQYMGvrZt2/rOOuusdPct1OtVdpVH+9euXTvf/v37fV988YWvYcOG/nLqfb1r164U61D/r5tuuslXoEABX4sWLXxnnHGG+/uGG25w6/Ds27fP9Y3RczpXp512mu+6667zPfroo8f0/Qr3OEWzr1NaxzLUdiLZH/UP6ty5s69gwYJuneorlNl9D0XfV8GfaZk4caKb/8wzz6SYH4/71a9fP/cdFEifZ71W50Xr1XftHXfc4fofpkXfUVq+b9++vkh43yvqcxeK+m7p+bfeesv1g1SfvuDPRVpCHfdofOb1faNjV7p0aXd89fzYsWPdedQUaPv27W473veqJp2bAQMG+Jc5evSoO59aRt+t+m7Tb4PmqS/joUOHXHkLFSrkfgO98geuIxSdb73H9L7Re0b7oH168skn/cv8+eef7jn9ruj7qF69er5ixYr5XnnllYj2Q+9BlUm/MSpn4O9KqLJr0rEL7qMXTplD9XsK5zgH9kl77LHH3Pes1q/XtGnTxt/HNr3jnd75Sou3fa1Ln0/95hYtWtR39tlnu33waP+0rD77gbRtlUnlTovW75W7Y8eO7jXqI6nPUUbKE+4+h/t+Cj6He/bsce+FYcOGHbMv+r3TNsM5N5G8F4Jddtll7jMd3Nda36n6vAce83CPb0b3M7P7Eo8I9HKZTZs2uYsh/cjqTa8P18033+z77rvv3Icq2Lx589wFkT4wukDyLFu2zB+4hJuURNtWUHnjjTf65+lLXh+ywC+sLVu2+D7++OOIlglFFwX6Mti4cWNE6woO9F588UVfvnz53AWJvnQ9ofZbHd/1Q6l9PHjwoJt35MgRX8+ePX1lypTxbd26NaLj45VFXzIrVqzwl3f37t3unOhC8vrrr/efG3056Qf/oosu8kUz0At3OyNHjvTlyZPHlTvwIlY/OvpSTsv06dPdawPPhbftK6+80v/DG3wjIi3e63VMvR8RBQneBWaPHj38wZrWq/f60KFDU6xDF9E6/yqfZ+bMmW4dffr0SXHs9CP4008/+eeNHz/e/VAFBwXhHqdoB3qpHcvU3gvh7s+tt97qfhADAxMFEZnZ90gCvd69e7v5uhkRvA/xtl9KhqEgLjULFy50waAu5rSNOnXq+IYMGeL7448/jlnWu2GkYCfagZ72RTdHLrzwQnfRpGN2//33p/gOTE2479tIPvP6PtVxV5DgBc4qy2233eaOVWCgp9+y1q1bu0Bl5cqVKW7+6bM8ZswY9/iDDz5wxyEwuYReq3MUeNGpG1u6sA6HXt+yZUt3o2/x4sX++frufPfdd93f+v1RkKpzu2HDBv8yDzzwgCuPbkBFsh/ecVM5g4Uqe3AylnDKHOriOZLy6f2gz8LTTz/tn6frDpVDv7HplTmS8xWKtq/3jz6ngTe79busG7geXddoG6+99lqK10+dOtXN13dFJPQevfvuu932A3//wy1POPsc7vspGgFQaucmkvdCqO8BlWHChAkhb9andcxTO76Z2c/kTOxLPCLQyyX0JtddPL1Jddfoiiuu8H3yySfuLk1adIdPd0vSylyZVqC3fv16V8Olu7m6qNDdKAUIHn3xpxeUhLNMuMJZl3cR9Msvv/juuusud0EffFcstf2+9NJLQ94N1IWJLmhGjRoV0fHxyjJ8+PBjtq8vMT0X+EUk/fv3d+dZAWa0Ar1wtqMLMb1W75lg+lHSOgIvVvVlqlphZfHy9l8/coHBk7ftwIuPSHivDw4OdddU8/XDGkh3DvWl79F51AWuPi/Brr32WhcYKuj2luvVq9cxy6k2IjAoiOQ4RTvQS+1YBm8nkv3xllWtZzDdpc7ovqcV6I0ePdq9X/Qa1azmzZvXBWU5db9Su4DRhYW2r+9grUPvzdWrV/syK71AT999upkXWBZdkOo1AwcOTHf9qb1vM/OZ1x17zQ++c6+bTjpPgYGebsRo2TfeeCNk7YEuikU3PbWcsoymJZJAz7toDQ4UAum3N1TWS72P1GJDNTKR7EdmA71wyhzq4jmS8nmBXvANZd1M6NSpU7pljuR8haLtq3VN8I0K3djTOv/66y//vKZNm7opkG546AZEYCuO1Oi748cff/RNnjzZvcefeuqpYwKucMsTzj6H+37KykAvkvdCMK+2VNeoga655pqQxzyc45uZ/ZyZiX2JR/TRyyX279/v2iGrnbKyWV1//fUugUF6/UDUF0J9P9TeOxI7duywq666ym1Tbc4rVqzotrV27VrXB8CjzG6DBw92fQXUBlttpPV3YLnCWSZckaxLx0ht3dUvQccqHOq/VqVKFfvqq6/c4/93M8VN6pOifgmRHB9PqD4RXv+k4H596nOh86w+ZVWrVrVoCGc76mOgPkN6ryiroPbZOwZaRtT3Q/2Wfv75Z3cudu7c6fa/ZMmSrg28kgEF77/6ijRp0iRTZVcfpkA6R0WKFDmm/5Tmf/PNN/7HytiqPgnq0xdMfRTfeecd135fZdRy6osZTO+vH3/80f84kuMUbeEeS5Uxkv1JbVn1afD6T0Vz39WnTevQutS/WH2JHn744Ry/X4G8PiNeH8A333zTfR//+++/rq9gVtL+BZelf//+9sknn9hzzz1nQ4cOTbUva2oy+5n3xhwLPh/q21OzZs1jvodFffy88+GdE6+/oMZLVN9N7cupp55qV1xxhetrpCkzWUV1jqR169apLuPtS/D3ivZbx0b9EiPZj0jPRUbKHEqk5dNxDu7bpt8R9YMKR2bPl16n93IgL1GNzomXyKlPnz7Wu3dvd1x0jtQ/V/1/NT+9vmBKFPXggw+6vqP169d3/dh1fCT4fR5OecLZ53DfT1kpM+9VHQP1sX7kkUdcfzwlA9T3qcZ9vPHGG1Mc80iObyz2JR4R6OUS6kCqDtlKTz5+/Hi76KKL3DwNP6Dsa+oEG6pz8YEDB9LtzBvKAw884L5cdHETGCAoA96ff/7pf6wPrC6YP/jgAxs3bpwNHDjQdTJ+8cUXXfrncJcJVyTrUsdoZSzThXy4gZ4CalFim2BaR+PGjSM6Pp7UOgAHpz73vty9cxct4WzH23dd9HpfvIF0fNWRX8Gh0uCfcMIJ7vgGfonrXAR+oYpek1rH98yUXYF3qPmBx00XoRLqM+DNUzDgCZWcJnheuMcpK4R7LL39Dmd/vGUVOGfXvgdm3dTNESXRUfIO/UCn9eMb7/vlUdIffVdPmDDBXfjogkYXe0rUEphVNLvpYlPHePPmzS6xRrii8ZnPyPlQohrveyr4fOhCTQGzbua88cYb9vXXX7shCvR51nF+7bXXQr42Pd73R1q/m+l9r3jfKeHuR2YvOMMpcyiRli+17+Jwf6sye77S+twHfo/rJuz999/vsowqeHr55ZfdezS9DL5KwnT33Xe7m04PPfRQigRD06ZNO+Z9Hk55wtnncN9PoXiBZnDZvDKEm/Qus+9VL9DTkBpDhgxx339aZ+Axj/T4ZnQ/92fT5y67EOjlIroDosxumlatWuWCkffee899melHu0ePHi7oC8z6pJoQ3b2IlO6MK3gMrgXysq4FB0BeIKWLGpVDX2LKLOl9+MJZJlzhruvxxx93QeE999zj7kKHU2OgL2X9aKWXujqS4yOZCXSyiy7iVE5lCH3sscdSXU7jd+mmw6BBg1Icb9VUaIqnffdqCpRxNZg3Txfe3o9BqCA9eLy0cI9TVgj3WHr7Hc7+eMvqrnew4HlZse/a/rPPPus+x6NGjXIXaGktG6/7pTLp4kbfybqwUwDkfTeFqlWMBQV42s9QF+xpUcuIzH7mA8+HavE8ygaoYD8wE7JX4/m///3P3QRIi27oqZWHJl3w6T3kve6WW25JtTyp8VoP6Bym1qIi8HslOHuv5nnBfCT7kRnhlDmUrCpfWsc7nPOVmrQ+94E3UBRsqUWPAjx9nhVchZO11msRoKAlnN/1cMuT3j6H+34KRfuqlgjbtm1LMd/7XAVn803t3GT2vaCaXX1/6lgriFNgqxtLgcc80uOb0f2snU2fu+ySss4YuYYCDN090ZeKmtRcfPHF7gOmD5bupnhUba4vI9UCBtJdEaVtTo2GLVBa88C7Jxq7STVYgbRMIFXZ68OuYMkb6DicZcIV6boU3OlLVYPOKkBO646RqLmH9lHpgoPpLpH3JRPu8YkWLx2y/s8qujDV+0hf0Fu2bAmZHlr7q+X0Y6Ev10AjRozI8JAZWUXn6ayzznK1v7t27fLPV2p3NaXT3V79yOrHVYG77kZqEO/AfVaTn4wcp1iey0j2R8vqx1ADgXt3Qr2gQHdZM7rvkeynaovU3FDvocDzFI/7FYoCPDXr1M0l7YfWr8BI3z1pBXlq3qRjpBrAaFJ5g6m2Uk1mdZMsVG14WqLxmdfxLVSokI0ZMybFfLXOCDw/0q1bNxcMjhw5MuRx934Hgs+L1q/mpRIYgKqZnI51OLp06eKGJNC+aSiIQN57Q83xtN9qBhtITfl1LhXgR7IfmRVOmUPJqvKldrzDPV+pURNHr5mjaF8VzOk7PLh7hJppKqjSPup3O5zxOPV7IYHvc703tY2MliecfQ73/ZQaBTmBXRZEv3mhagJTOzfReC/oGOv7Xt97qr0LPuaRHt+M7me3bPrcZRdq9GAtW7Z0kz5cutgJ/KJX4KIxxnr27On6pGg5fenp4mLAgAGun1so/fr1c18+ugDr1KmTG/9FzQ6uu+4619zAoyakaoKlCyr1U1NNo+7OK8D0miuGs0y4MrIuBXj6EdQyqtnTBX5q/QN1vNQPT18U2lfVCKrZki6S1L9FAbOaJYV7fKJFX/i6C6ZgXncqs8rrr7/umtCpiaruNOpOsd4veg+peYXuLuqumsY6U/M7HUfdPVOfRl0MKvCON9on9YnQe1/l1gWr15RH7wWP5qlPlQKJm266yd080LhDOt56j0V6nNIapyk7zmUk+/PKK6/Y2Wef7YJi/ThrWb3f9b2hZQP3Jdx9j2Q/tfzw4cPd51uDfGtQ8Hjdr1DUokItLPR9GskYmKopU/N7fUfp+zstat6lscwC+2QF9t0JHBNQ+6fvPL3n1RJE319q+aG77hn5forGZ14XedrH2267zV3c6Qadvr910RU4Fpuo/5/Oqb6HdcGsliraDx0v/cZpLC2d24kTJ7rgXMdd50vBv95v+m3Qd7FH4xpqv1V+9eNNaxw9bVvrVfMuff+rGaACY91Q1c1R3QTUfmv7+k3R74Am9VN75pln3Hq9Wulw9yOzwilzaq/LivKldrzDPV+p0e+tfnt1s0K/w/o9VlNi9cMPHhdY69dyar6nIErHJJybEWpmrUDsvvvuc/P0udZr9T2QkfKEs8/hvp9So5pCfXfq/Ot6RDec9RkL1Z87tXMTjfeCvoN0HaauLaGOeaTHN6P7WTKbPnfZhUAPfvqR15s/kH48dQGjH35d4CgY0Q/zq6++6hIgiL5wgvugaD26iNCdar1Gd1JUY6UBK5WIxOPVfqlKXoMba126GAn8EQ1nmXCFsy5ddGl/9GH3aP/1WHd/tE/6ogm13/L888+7L2Dtqy5W9To1K1q4cKG/yVG4xydUWQI7bYdqk69mGnqNkpB4dIdQ8wKTFqhJRHAtZqh5kWxHX9LaL71XtO/6UlTNyKWXXuq+kL128uoTqQtt3TzQXUmtx7tg1Y9HetsOV2qv15d3qPb1qj1RU45A+mHRj69qgXQOFeApQFBgENiETF/+CvIVSOgCSc35dENEF9W6GA1s6x/ucQp1PkKdy8wci9S2E8n+eMvqx0/JUVQ7pQsUL3AKDF7C3fdQ+6ntaF6oGiUlmbrrrrtcTZhuVimgiMf9CkWBpKZI6XtBx8P7Lk6L7lqrz59Hr9Odce/uuIJU7zOh7ySVX/uh97yCPtWk6TXh1MCFOu7R+MwrWFQgreatKpvWp9YW2v/gJocK5tViRcuq1kTnUd9ZOndewot7773XfRfr/OtOv86nLpa9QMej5nv63dOx0G+HLrjT+v3RhbuCUP1e6HdG+6PB3K+++uoUvyn6vtHFvbatYFi/tXq/BAYd4eyHd9xC9V/UxXHw+0NNb3X8A4PscMqs90jwYNPhlk8X8WoxFEw3XYOD/dSOd7jnKxRtX9tSgKKbFt9//71rkaGgJbgLhUc3dxTo6YI/1GD3wVQeBRz6rOh/BUH6WzcpfvjhhxRNKMMtT7j7HO77KdQ5VPCmz6XOof7XsdY1jJpaB/9OpvVZCPe9kBp996oVlb5/dPMm+LonkuOb2f08J5P7Ek/yKPVmrAsBAEg8ushQbVZWNjONhUTdr5xISRF0waebbwomgWhRXzHdSFDg0bZtWw4sciT66AEAMkXNX4KzTapvr2rP1f8np0rU/cqpghMpiGotVHvI+UA0qQ5ENeBq0ZGR1kNAvKDpJoAcRc3r0muPryZ3wc02ElG8HAtdaKuZl5qDqSmPMiO+8MILrllkYHKnnCZR9yunUs2Kau3UB0m1eBrqQf2X1JdTffaAzNKNnS+++MI12VS/WnWnyAlZr4HU0HQTQI6itvjq15AWdZ6OdIzFnCiejoWyw6l/iIZjUX809WNQc7pIkovEo0Tdr5xKwZ1qVNVsVn1klVxB/XGAaFBW3TvuuMP1cVMCFN1UAHIyAj0AAAAASDD00QMAAACABEOgBwAAAAAJhkAPAIAAmzZtckMoaCBijcf33HPPufkae0ljVWmMMc1fuXKlG/A4cCzJcGicK71+1qxZHPcM9pvU2GYaxDjw/AAAUiLQA4BcQkGJLozTmzRAfG525513uuQrCxYssAMHDrjHGmqhR48ebvBpJWzQfKVeP3r0qBugPRLJycnudVk9jK2yB+p8arD37KD9euqpp1yWVwXDylT66KOPRnx80tOnTx83uLcGMvbOT3rZaXv37u0GetZg4a1bt7bPP//8mGyLqX0e+vfvH9XyA0B2IdADgFziySefdBfGoab333/fBR/lypWzypUrW26mGrfzzz/fqlat6i70lV79999/ty1bttjll1/uMvJpvjzzzDPHjLWXnnbt2rmBvs855xzLStkVUHruueceGzp0qI0cOdIdKw2FoMDvlltuiep2VBOq81OlShX/+UnNkiVLrGnTpu54T58+3f755x8bNWqUvfLKK7Z7927/cjpGOlYDBw485rMxYsSIqJYfALILgR4A5BIaHiBUjYWaIN54441WsGBB++ijj6xUqVKWWx06dMh27dp1zPAJClwkeL6Oab58+SLahgKT9AKUnEbvIY0x+MADD7jBy9WcVcHYww8/bG+++aYtXLgwKts5ePCgJSUlhTW8hQI3DS/SsmVLe+utt1xNo1532mmn2dSpU13tXjifEc0DgJyIby8AyMUU1OjC/L///nMX6meeeeYxy7z77rvuYrl48eIuCNS4ZcEX7rqwV9NQNRM866yz3OMnnnjCPacxz3r27GmVKlVyTfoaNWrkan10IR5O+dR0ThfpWqf+12DlgbVo4a5fNTj333+/1alTx13wV69e3e6++243MLo89thj/gBCfwde7F988cVuvgZQDw4EgvvopVfm1PropVc++fjjj91r1axUNVO1atVyNYyqJVy+fLl/uZdeeskuueSSY8qsoEtUUzVo0CDX/FSvVzNL1bxt3LgxRZlUFk3p0cDSqkFU37lA3hiOH374YbrrSO886pxofvD5UfAXyrRp09wxueuuu9LdNgAkJB8AIFc6evSo78ILL1S7Pt+tt94acpn+/fv7ChYs6Hv66ad9f//9t2/dunW+m2++2Ve0aFHf4sWL/csVKlTI16FDB1/nzp19S5cu9f3555++Tz75xLdx40ZfhQoVfE2aNPHNnTvXt23bNt/YsWN9RYoU8fXo0SPN8u3cudPXoEEDX40aNXxTp051r125cqVv2LBhbh0S7vp3797ta9iwoa9evXq+r776yq179uzZvhNPPNF35pln+g4fPuxLTk72/ffff+54DBw40M3zJu2L5n/33Xcp5t91111u3yMp84wZM9y6vv3224jKJx9++KF77bXXXut74oknfJs3b3bHu1GjRr769eu7cyral1Bl1nzp06ePr3z58r5vvvnGl5SU5FuzZo3vtdde8/Xr1y/FOdB+aErPlVde6cuTJ4/v4MGDxzxXvHhx995ISzjnMbXzk5r77rvPLfvxxx/7zj77bN/xxx/vK1OmjK9Lly7umAXas2ePW7Z06dJuOZW5RYsWvjfeeCPdfQeAeEWgBwC51AMPPOAubs866yzfoUOHjnl+2bJl7uJ98ODBKeYrmGjcuLHv4osv9s9TsFOsWDEXsAS6/fbbffny5fOtXr06xfxHH33Ubfunn35Ks3x58+b1LVmyJNVlwl2/AgMtp30KtHDhQrfcBx984B7v37/fPR40aFCK5T799FM3/4cffkgxv2/fvikCvXDKHCrQC7d8XqB30003pVhu8uTJbr4Ct/TKLApGr7rqKl96jhw54qb0tG/f3nfccceFfK569eouaEpLuOcxtfMTigJELatg8cUXX3TB46JFi3ynnXaaC+RWrVrlX1YBZLdu3XyzZs1yQbaC87vvvtu9XmUDgJyIppsAkAt98MEHrmmlEo5MmjTJChQocMwyn376qUtSoQQkgdRnqW3btvbtt9+mmK95at4ZaObMmS5TZe3atVPM99ap51Pz2WefWZMmTdyUmnDXrz5ZJ510kjVs2DDFckrUUaZMGTd0QjSEU+ZQIi2f15TU423vr7/+Cmt7jRs3tilTprhkKWm9Rv0Pw+2DmFqfQ81PLyFMZt4nqfGafKo5qDJ16jhqG0o8pKbKgUlW1CRU/VPPPvtsK1mypNWrV8+effZZu+6661xSmaVLl0a8fQCINQI9AMhllIlQyVcKFy7s+nxpvLhQNIyAaOw4LVuoUCGXsEXT888/7y6WlbzEoyyIwbZv3+76XAWrWLGi+3/r1q2plvPff/8Nuc6MrF/78ttvv7n9CNwXBbhax7Zt2ywawilzKJGWL3ifvQBb/QPDMWbMGLv66qtdXzcFV9WqVXPBUHAfvXCVLl3a9UFUdstgO3fudEFWWjLzPklN2bJl3f/B2U3VH1HnaN68eemuo2vXru7/7BqiAgCiiUAPAHIRXVAr+cq+fftcivlmzZqluqx3ca6hBRTU6UJer9OkBBi6qFcw4glVK6gAwMtYGcibl1YAoKEeNHh5WsJdv/5XohntR+C+aHw87ceECRMsGsIpcyiRli+12rNwh1LQ9l599VUXQCnwV8IS1fIqKFJSlUiphlDb1vh2gTScgbJk6vm0ZOZ9khqvljPUsQo3k6a3XHYNUQEA0USgBwC5hAauVlO4tWvXWt++fV2TtrR07NjRn1ExtcGk09OmTRtbvHixG7Q6kNbpNfdMjbJ7/vrrry7QzOz6O3XqZL/88ov9/fffIfcjWin0wylzKFlRPi9DZWCtazA1y1RApGyfGgfvzz//zFCtnmq+FFB98sknIc+Dl30zK94nqVHWUd2I+PHHH1PMX7NmjW3YsMFatGiR7jrUfFnOOOOMiLcPADEX606CAIDsoQya+tpv06ZNmtkKA91zzz2+woUL+5555hmXGVFZFVesWOF79tlnXcZJjxKSKDFJsLVr17pMhs2bN/f99ttvLumFEosoccsll1yS5raVPKN27dq+OnXquAQmeu369et9Tz75pG/cuHERrX/79u0uK6WyUyqrpTJN7tq1y2W27N27t0tcEo1kLOGUOVQylnDL5yVjmT9/fopybN261c3Xdjw6X/nz53cJYoKT7Zx33nnuOG3YsMElW9E5Pf30013ZA5Ov6LGmcKicOu5ffvmle38psYmyXIaT9CXc8xhJMhZ55JFH3Pl566233Dq1n8piWrJkyRSJX5T0ZcSIEb7ly5e797iOy4MPPuiSEd14441hbQsA4g2BHgDkEuecc467SFZWSGU4TG0aP358ite9++677uJY2Qt10ayMjUpdrwAmvUBPlMGwe/fuvhIlSrht16xZ02XyDJWKP5gCGA0FUKVKFVc2vVYBa2B2z3DXr8Dp/vvvd4GL1qXAolWrVr7XX3/dd+DAgagEeuGUOVSgF275Ign05Pnnn/dVq1bNBSxapzdcwJw5c1wAVrVqVV+BAgVcWZXJU8NnZGR4BVFw9/DDD7ssm9qW1jlgwAB/2dMTznmMNNCTl156yb1nFfQqcNTwCgrogo+fyn7SSSe586mbG82aNfONGTPGP2QFAOQ0efRPrGsVAQBZT1kIw/nKV3O+1PqApbVuvSZaTSBj3cRV+xG4Lzpu2sfg5qrqz6bnws1MGbiujBzn1MrhlTu1dWZ0m17mykj2LxbnBwBwLAI9AAAAAEgw3A4DAAAAgARDoAcAAAAACYZADwAAAAASDIEeAAAAACQYAj0AAAAASDDH5mdGVCn19qZNm6xYsWIRp9EGAAAAgMDhcvbs2WOVK1dOd5gZAr0spiCvWrVqWb0ZAAAAALnEhg0brGrVqmkuQ6CXxVST552M4sWLZ/XmErpmdOfOnVaqVCkGyeU85Hq7du2yGjVq2Lp166xkyZK5/njwvZS78fsQe5yD+MB5yB3nISkpyVUieTFGWgj0spjXXFNBHoFe5j40R44ccccwKz404DzktM+D8L0S+/PA91LscR5ij3MQHzgPues85AmjSxhXzAAAAACQYAj0AAAAACDBEOgBAAAAQIKhj16ctOU9dOhQrIsR98fo8OHDduDAAfroJcB5KFCggOXLly+qZQMAAMD/j0AvxhTgrVmzxp9cAamPGeJlMWI8wsQ4D8oWWbFiRc4nAABAFiDQi/FF8z///ONqNpQmlWySaR8rZTDKnz8/gUEOPw9ax759++zff/91jytVqhTlUgIAAIBAL4Z0wawLXo1sX7RoUd6NaSDQS6zzUKRIEfe/gr3y5cvTjBMAACDKSMYSQ0ePHnX/FyxYMJbFAGLCu7mhPn8AAACILgK9OECfM+RGvO8BAACyDoEe4srBgwdtxYoV/trOjC6T02if/vzzT/v777/987Zt22YrV660Xbt22datW1M8F876Eu0YAQAAIHwEeoiYEsgoiNCk4ESp9qNl+fLlduKJJ7qsjqJ1azuBWUmDl8npJkyY4PqpXXDBBfboo4+6eTfccIPVqlXLLrnkEps+fbo9/vjjdtNNN4W9zqw6RqHOBwAAAOIPyVji0AXDPs/W7U0f3DGi5QcMGGATJ0606tWru/5VGzZssIsuushef/11K1euXKbKUrhwYatfv75L9iFLly61Zs2a2Y4dO+z4448PuUxO99BDD9nw4cPtzjvvdI9XrVplb731lv31118u2JN169a5zKzhyqpjpPPRvHlzdz5KlSoV1XUDAAAgehLjShnZThf73333nft7/fr1dtZZZ9ldd91l77//vpunJoMKAIsVK2ZlypQJuQ41Sdy/f3+K9PoKbKZMmeJepyBSAY788ccfLlNj6dKlUyyjmiUFRgqCjjvuuBTrV7NHZTTVcp6NGzdaoUKFrGzZsmHv6+bNm13CHG07WDj7mdp2NYaigjm9XgOIq6ZM5s6dm6L5pXTu3Nkf6IZTvsBjFE5ZROdCx1sBoo6ram41zp0XLAafj+LFi7tJxxgAAADxhaabyDTV7F1xxRX+wE/BnoI3BYNVqlSx8847z7Zs2eJfXin1W7du7YIzLXPCCSfYjBkzjmlyqOX69evn5l9zzTV26aWX2tNPP51iGY092KVLFxs1alSKMv3yyy/WoEED/3a//fZbq1evnp1yyinutY0aNbL58+enuV/ffPONW0ft2rVd8KNaS2/st3D2M73tKmhS2RXQqcmm/tY0ePBg97z3WFOrVq2OabqZVvlCNd1M7xgsWLDAzVcNY4UKFdx+KXj8+OOP/eftvvvu858PlUvnAwAAAPGHQA9RoYBCNU6q6bnuuuvssccec4GBapv+++8/u+WWW/zL6jkNEq9kI6rNmjVrlr/mKpCCp0mTJvlrudRsMFRgcfXVV9v48eNTzNPjM844w+rUqeP6Eaqvm4IpJTXR1LNnT+vWrZvt3bs35P6olrBTp04ugN2zZ497Td++fV0toYSzn+ltt27dum6/VRs3ZswYf7/HN998073ee6xJffYiKV+wSI6B9k2JXzZt2mQPPPCA9e7d29Xm6Xx8+OGH/vOhchHoAQAAxCcCPWSIBnrXhb6Cr1deecXeffddFzi89tprrqbIq30qWbKkjRgxwj755BMXOEhSUpKrKfLGD1TNntc/LSMU6KkGa9GiRf7mlEpwolonefXVV+2kk06ypk2buiBGQZKCHgWa8+bNC7lO7ZOaPw4dOtTVGsqFF17oaiIlnP3MyHbDlV75gkVSFgWD6uMnOobqj6fmuQAAAMg56KOHDFGAp6Z76ltWtWpVe/755+3GG2+0yy67zBo3bpxi2SZNmrj/V69e7fpz3XvvvdaxY0dXo6VMkwpQ9NgLWCKlpp8tW7Z0tXgKZNSkcfv27a62S5YtW+YCQdWABapRo4YLOkNREKt1pTbWm/Ylvf3MyHbDlV75gkVSFtXcebx+gao1BAAAQM5BoIdMJ2MJpIQpSuoRyHtctGhR979qlpSEZPbs2a7f2B133OGaLn7xxRcZPhuqeVJN1BNPPOECPgWQXrIRBaOq6Zo6dWrY61OyEtVapiac/czIdqNVvmBZWRYAAADEH5puIqpUy/TTTz+5jJKer7/+2gVGShgi6u+lPnrK1KnEIxpK4MsvvwxZy+U1IdRr0nL55Ze7fmdaj5KHeM02RdtRP0A9H0iZJVMbD+7MM8+0H374wfW7C3TkyJGw9zMj2w1XeuULFq2yhHs+AAAAEFvU6CGqbr31VpcB86qrrnJNNNW3S5kaldTDS/V/7bXXumCoXbt2LnBQbV7Dhg1dqv5galqoZdQHUFktUxvCQLV3qsXT9r3hCDxKkDJ27Fj3/MMPP+wyZf722282evRoV6MYajy4m2++2fXD69Chgz344IOulk5l0GOtO5z9zMh2w5Ve+YJFqyyB50NZPkuUKMHwCjHU/ckZlr/w/9UgAwCA+BhzOl5Qo4eIKUjQkAqhqE/Xjz/+6IK2Pn362IsvvmhDhgxxk0cBisZmU22ext4rX768TZ8+PeRA3wqa3nnnHfvss89cnztleUxtMPBevXq5sfQUBKlmLbBMqn1TP0AlTLntttvs559/tvfeey/VAEfl12s0WPvAgQNdWU899VR/EBXOfoa7Xe1L4Bh52gevVtCjYxQ4YHp65Qs+RuGURcGiXhPYV1Kv1zyvJk/nQwGjzkf37t3JugkAABCn8vh8Pl+sC5HI1BxRtR67d+8+psbqwIEDtmbNGpc90buQRmh6m6pZogKPcBOQIL7PA+//jNu1a5cL0NsOnEiNHgAAcVSjl5yc7DKWK8N8RhMNZjS2CEaNHgAAAAAkGAI9AAAAAEgwBHoAAAAAkGAI9AAAAAAgwTC8Qho++ugjN6h3cMbJfv36ZfV5AQAAAIAMo0YvDTNmzLCvvvrKKlas6J9SG8cNAAAAAOJFrq/R27t3r02cONE/zMGll17qH/Ba6tWr5wbCBgAAAICcIlfX6O3bt89OP/10NwB0gQIFbM6cOdaqVSvbs2ePf5m//vrLHnroITcw9Lx582JaXgAAAAAIR66u0VuwYIGryfv1119doCd///235cuXz/2tAaG92r3ff//dBg0aZLfeequNGjUq1XUePHjQTYGDGgIAAABAdsrVNXrVqlWzo0eP2nPPPWfbt29386pWrWpFixZ1fz/wwAM2a9Yse+SRR+yNN96wzz77zEaPHm3ffPNNquscMWKEG63em7QNxMbNN99sQ4YMSZjDP3z4cDv11FNdE+PffvvNtm7daldccYVrXtysWTNXQ12zZk1btGhRrj1GAAAA+D+5ukZPF8WffvqpPfXUUzZ48GB3wXzTTTfZHXfc4Z7XBXWg8847zwWCP/zwg7Vt2zbkOgcMGGD33ntvihq9iIO9oV0tWw2dHNHi6rM4adIk69+/v912223++b/88ovr4zh//nwrV66cZadevXpZjRo1XDNbz7///msFCxa0REkM9Pjjj9u0adOsSpUqVrlyZbv//vtt06ZNNnXqVDvuuOMsOTnZ1q1bl6JGOT1ZdYxCnQ8AAABkn1wd6MmFF17oJl0cq8bummuucRfR3bp1C7n8kSNH0ryQLlSokJsS2bZt21wTV13EX3311Va8eHE3/8CBAy7QUC1pdtuyZUuKJDry6quvWv78ifEWVw1e3bp1XR/SwHlt2rSxBg0auMc+n881RdYQIOHKqmMU6nwAAAAg++TqppsrV660DRs2uL8VnHXt2tUqVKhgGzdudAHdjz/+mGJ5ZefcvHmztW/f3nK7c845xwV4qmVKy8cff+xqQhWMdOjQwdVMBVJQ+PDDD1vjxo1dEPPYY4/ZXXfdZf/73//8y6hvpGpXFegoec7tt9/ugs3AGkY1p33zzTddLa2mP//80x588EHX1FbGjRtnLVq0cMFQIL32+uuvD7u8oUyePNnOP/98O/HEE+3yyy+3P/74I8XzL774orVs2dLq169vl112mS1btiyi49SnTx8bNmyYe523f5p+/vln1+zYe6xjdO6557r+pOGWL/AYhXsMrrvuOtecWbXgZ5xxhms2+swzz6R7PgAAAJB9cnWgt3//fndB27FjR7vzzjtdoKFELLoYViIW9V06++yzXfNELdOzZ083T7UouZ2S16jPmBLTqPlgKK+88ooL2tQPTMGD+pOppnTmzJn+ZRQwvPTSS+5/BS1LlixxgVFgIKemsN9++60LOPTc+vXrrXPnzv7nFRQqkFKz0e+++85N1atXd80Svb6XCs7VtPSnn37yv+7QoUMuGFHQGm55g6k8qgXW++ODDz5wgZya73qeeOIJFxD17dvXJkyY4ILjs846y5Ut3OOkQFjPKVD09k+TgmMFqd7jL7744pimm+mVL/AYhXsMVFunwFOflddee801IVUwrua8aZ0PAAAAZJ/EaNeWQaeccorLuPn111+7Jm+q9bjgggv8fZY0X0MqKLmFgjsFJOp3hP/To0cPN+zE0KFDXRPAQKoRHThwoEti06VLFzevYcOGtmLFCnv22WddgK2ARH8rWFRtqrz99tsuqAukQepLly7t1lmnTh177733rGTJkrZ8+XJXS1W2bFkrUqSIC6JUexSKmjOqX+X48eNdoCUKjBTsd+/ePazyBlOgqBox7b8COWnSpIlbnxw+fNgeffRRl6BHx0oUGKmPp/ZZtZfhbFf9HUuVKuXel4H7p1poHQdv3n///RdR+YJFcgz0t9YrCjhV2/3VV1+54C6c8wEAAICslasDPe9iWc3TUtO8eXM34Viq9VTTTQXH99xzT4rnFITt2LHDBRhqyqcmk5qUnEZBm6xdu9aNWdi6dWv/6woXLuyaZwZSjaGCJY1zqNonJR3RuvR6BXrhUs2WagdVc6gaSQV9qhlUQKL+bumVN5j2cdeuXa6PZ6C8ef+volw3D/R6NacMfE41iKq5DPc4ZVR65Qu1fLhlOemkk1I8rlixomvyDAAAgPiQ6wM9ZE67du1cTZmaAyo48Cgxi6hpZO3atVO+6f5f8g9vmeDkNcGPL7roItf/bOTIkS6DqYI0NWOMJLukqAmimuF++eWXLthS8h3VRIVb3mCqsRPVXoWS1v55z2Vku+FKr3yplTecsnhjTQYK7v8IAACA2CHQQ6apVk+1cOqX5dFQFQoOVKulYDCUE044wdUuLV26NEUfLiUdUd9IrzZPzWs//PBDF+xpnatWrfIHMYGBR3qBhrJAqgZPNXnqA3j88cf7a7vCKW8wJYfRazSchJqUBlOwpP1T+QMDJ9XmNWrUKMPbDVd65QsWzbKEcz4AAACQdXJ1MhZER9OmTV0fNNW4eTRYfO/evV0iktmzZ7t5Cs5UmzZmzBh/4KXXKcGNmmQqMFCzysCskGoyqOac6v8lO3fudFk3g2lIjOBsl6FoOAiNnag+hUq6o9rBcMsbTK+54YYbXCISNf30kpsoSY1obDs9r/37559/3Lx33nnHZctUJs2Mbjdc6ZUv1PLRKku45wMAAABZg0APUaHgYd++fSnmKeHItdde6/rwKZmIgjYltAnss6ZlFBBpWAslXVEQpto8LwBTkKcEJgpWFDyo6aZqw4oWLZpiW7feeqsLZpS4JK10/qrB02vnzp3r+uxFWt5gCkzVtFT9OFV+JfhRzZhHyWr0WEl8lDhFzVvHjh3rkqJkZrvhSq98waJVlnDPBwAAALJGHh/tq7KUElmopmT37t3+gcUD+0SpmZyaJCqgySnU7FHj3yk4C6RaK/WbUzAW2IdLy6omSRf9qfU7U42eAj4dBwUjGgZAwV1gRkg14yxfvrxbRkMsKLtjYMCnt7LKtnfvXqtSpYqr/dP2ghOJaHgAZdtMLSNkOOUNpgyX2l7wMfGoTHov6PnUkqGktV29fxRIBw6GrjEddSwUQHr7r+EVtExwv8DUyqfthTpGqZVF21DSFR33wNcoiYvOkc5PaufDC95z+vs/HijJjgLxtgMnWv7CKW96AACA6Jo+uGPYyyppoK6LdJ2U2jVfVsUWwQj0slgiBnrRpCaCCkBU06Ysns8//7zLjKnaoMCMmgoaFEgo6NByiI1ongfe/5kP9PTZ8QJ9ZL+s/jEH5yGn4LMQHzgP8SE5jgI9krEgppQ9U/3I1HdOtYGqEdLA25EMmwAAAAAgJQI9xJTudnzyySeudkeBnu5QAAAAAMgcAj3EBTVdza3NVwEAAIBoo1MBAAAAACQYAr04QOJT5Ea87wEAALIOTTdjyBuCQKnvixQpEsuiANnOG3cxeNgFhK/7kzMYXgEAELOhBBDfCPRiefDz53fjkW3dutVd7JKeO3UMr5A450HrUJCncfo0NEDgmIsAAACIDgK9GNKFsga31lh6GugaaQcHGpdEwTDj6CXGeVCQV7FixaiVDQAAAP8/Ar0YK1iwoNWtW9c130TqFFxoYEgNv0DNZ84/D6rBpiYPAAAg6xDoxQFdMDO0QPoBhpr76TgR6MUO5wEAACBnIOsmAAAAACSYhAz0HnjgAZsxY0aWrPuhhx6yzz77LEvWDQAAAADRkJCB3tSpU23lypVZsu4vvvjCli5dmiXrBgAAAIBooI9ehIYNG2bVq1ePysEHAAAAgKyQ0IHe+++/b7Nnz3aZLXv37m3169dP0byzTZs2tn37dps/f74bF+ymm26yevXq2bhx49zrypcvb3feeadVqFDB/zrNP3r0qDVq1ChGewUAAAAAubDppjzyyCM2duxYa9CggRujrlmzZrZixYoUzTuvvvpq19+uYcOGrqlnixYtrFOnTvb111/baaedZvPmzbOzzz7bDRDtoekmAAAAgHiXsDV6qoWbNm2aS8Xfp08fu+CCC2zIkCH2wQcf+Jc588wzbcKECe7va665xsqUKWPHH3+8vfvuu27e5Zdf7ubNnTvXWrVqFdZ2Dx486CZPUlJS1PcNAAAAAHJljd4ll1ySYry17t272w8//JBimXPPPdf/93HHHeeaap5zzjn+eaVKlbKSJUvapk2bwt7uiBEj3GDS3lStWrVM7wsAAAAARCJhAz0FaYFKly7t+uMFCh6kPNTA5ZqnPnnhGjBggO3evds/bdiwIUPlBwAAAICMStimm+qXF2jt2rXZUrtWqFAhNwEAAABArORN5IybXpNL1ay99tprdtlll8W6WAAAAACQ5RK2Rk+ZNJs3b26nnnqqLVq0yMqWLWv9+/ePdbEAAAAAIMslZKD35JNPujHzihUrZgsXLnTj6CnxSoECBY5ZJtCoUaOscePGKea9/PLLbmgGDwOmAwAAAIh3eXw+ny/WhUhkGl5B2TfVfLR48eKxLk6OlZycbDt27HBJdQKzqYLzkBvt2rXLJZxqO3Ci5S9cNNbFAQAkkOmDO8a6CDlachZfs0YSW3DFDAAAAAAJhkAPAAAAABIMgR4AAAAAJBgCPQAAAABIMAR6AAAAAJBgEnJ4BQDIDT66v72VLFky1sXItcgGHB84D7HHOYgPnAcEo0YPAAAAABIMgR4AAAAAJBgCPQAAAABIMAR6AAAAAJBgCPQAAAAAIMGQdRMAcqjuT86w/IWLxroYAJBrTB/cMdZFAMJGjR4AAAAAJBgCPQAAAABIMAR6AAAAAJBgCPQAAAAAIMEQ6AEAAABAgsm1gd7JJ59sL7/8cqyLAQAAAABRl2sDvYMHD9qRI0diXQwAAAAAiLpcG+gBAAAAQKLK1YHexo0brXv37la5cmWrWbOmPffcc8c073zwwQfTXAYAAAAA4k2uDvSeeOIJa9eunS1evNgef/xx69+/v3388ccpmneOGDEizWUAAAAAIN7k6kCvU6dO1qdPHytfvrxdccUVdtNNN9lTTz0V8TKBFBwmJSWlmAAAAAAgO+XqQK9FixbHPF62bFnEywRSDWCJEiX8U7Vq1aJcagAAAABIW64O9PLkyZPicd68eS05OTniZQINGDDAdu/e7Z82bNgQ5VIDAAAAQNryWy62cOHCFI8XLFhgDRo0iHiZQIUKFXITAAAAAMRKrq7RmzJlir3//vuuX9306dPt1Vdftb59+0a8DAAAAADEk1xdo6ckK6+99pr17t3b8ufPb/369bNrrrkm4mUAAAAAIJ7k2kDv119/tXz58rnJ5/Md0xfPU7duXRs9enSaywAAAABAPMm1gV7BggX9f4cTwBHkAQAAAMgpcnUfPQAAAABIRLm2Ri+S5p0AAAAAkJMQ6IXZvBMAAAAAcgqabgIAAABAgqFGDwByqI/ub28lS5aMdTFyreTkZNuxY4eVLl3a8ublvinnIffiswDEJ36ZAAAAACDBEOgBAAAAQIIh0AMAAACABEOgBwAAAAAJhkAPAAAAABIMWTcBIIfq/uQMy1+4aKyLASCOTB/cMdZFABAnqNEDAAAAgARDoAcAAAAACYZADwAAAAASDIEeAAAAACQYAj0AAAAASDAEegAAAACQYBJueIW5c+fazz//bPnz57ezzjrLmjZtmuF1TZ061ebNm5diXoUKFezOO++MQkkBAAAAIGskTI1ecnKytW7d2u6++27bsGGDLVu2zD1+4IEHMrzOL774wqZMmWKFCxf2T4UKFYpquQEAAADAcnuN3oEDB2zy5Mm2Zs0aq1Wrll1yySVWtGhRy5Mnj40cOdJatWrlX7ZTp0528cUX2/XXX28NGzaMaH2eBg0a2IMPPpgt+wYAAAAAua5GT0FZ8+bNbfTo0bZ3716bNm2anXnmmbZnzx4X6AUGedKsWTP3/7p16yJen0evfeyxx+zFF1+0JUuWZPEeAgAAAEAuq9GbP3++rVq1ypKSkqxgwYJunmri8uXLF3L58ePHu6aWp512WobXlzdvXtu1a5drCnrvvffaPffc42oOU3Pw4EE3ebRuAAAAAMhOOapGr0qVKnbkyBF77bXX/AGUmlsGNrX0zJ492wYNGmTDhw+38uXLZ2h9CuyU3OWJJ55wQaOaeD7++OM2a9asVMs4YsQIK1GihH+qVq1alPYeAAAAABIw0DvhhBPso48+svfff99lvzzjjDPs9ddfP2a5hQsXWocOHey2226z++67L8Prq1evXorltU4Fh2kFegMGDLDdu3f7JyWGAQAAAIDslKOaboqSpWj677//7JNPPrFevXpZuXLl3DxZtGiRnXfeeXbdddfZM888k+n1hcruuX///lTXp6aiZOYEAAAAEEs5qkZv9erVtnnzZvf38ccfb1dddZVVqlTJn2xl8eLFLsjr2bOnjRo1KlPrU5NO9eELpKEW/vnnH2vXrl2W7B8AAAAA5LoaPfWju+iii6xJkyau2aUCsaNHj9qll17qsma2b9/eDZSuoC1wSISuXbuGTMiS1vqUxVOJV9Rfr1GjRrZ+/Xr77LPPrH///i6YBAAAAIB4laMCvVNPPdUNcaBhEJQd8/bbb3dj5RUpUsQ1p+zbt2/I16WWlTOt9cmPP/5o33//vWsO2rRpU5dts27dulm6jwAAAACQqwI9UQ1bt27djpmv4CwjA5untj7P2Wef7SYAAAAAyClyVB89AAAAAED6CPQAAAAAIMEQ6AEAAABAgiHQAwAAAIAEQ6AHAAAAAAkmx2XdBAD8n4/ub28lS5bkcMRIcnKy7dixw0qXLm1583LflPMAAPGFXyYAAAAASDAEegAAAACQYAj0AAAAACDBEOgBAAAAQIIh0AMAAACABEPWTQDIobo/OcPyFy4a62IAcWn64I6xLgIAxBQ1egAAAACQYAj0AAAAACDBEOgBAAAAQIIh0AMAAACABEOgBwAAAAAJhkAPAAAAABIMgV6YXn75ZWvQoIGNGDEia88IAAAAAGQS4+iFYcmSJf4Ab8uWLZk95gAAAACQ+DV6t912m6stO/HEE+2cc86xIUOG2P79+1Msc/DgQRs4cKCdeuqpdsEFF9gbb7xhPXv2tBdeeCHFcuPHj7fzzz/fLdejRw8XpKVl9uzZ1r17d7e8/v/hhx9SPL937163npdeeslKlSoVxb0GAAAAgAQO9AYNGmRTpkyxyZMnu7+nTZtmvXr1SrHMXXfd5YK4oUOHuoDv7bfftvfff9+2bdvmX2b48OEuSOzdu7e9/vrrdtppp1nLli1t+fLlIbe7efNmO++881yQp8Dx+uuvd68PXOftt99ubdu2tY4dO2bhEQAAAACABGu6WbVqVf/fqtmrVq2aNWzY0F555RUrVqyYay755ptv2tSpU+2iiy5yy02cONFq1Kjhf92OHTtcoDdr1ixr0aKFm6cAbtGiRfbcc8/ZmDFjjtnuihUr7OjRozZgwADLmzevNW3a1Dp16mTJycnueQWW8+bNswULFoS9L6p51ORJSkrK4FEBAAAAgBwc6K1du9aefPJJW7x4sQvYFHx58xs3bmzLli2zI0eOuGadnooVK1q9evX8jxWQKcC68cYbLU+ePObz+dz077//2sknnxxyuwrsypYt65p6qhlomzZtXJCZL18+W716tfXt29dmzJhhRYoUCXtf1Jfv4YcfztTxAAAAAIAcHeipL965555rZ555pg0bNszKly/vasFatWrlrxnTMqpxK1iwYIrXBgZgXp++d955x4oWLZpiueDHnhIlSrg+fHrNhx9+6JpparuTJk1yzUf37dtnV155pX95BZ7r1693zyn4VEAYTLWD9957r/+x9kXBIwAAAADkmkBv6dKltm7dOvv999/9Adl3332XYpnatWu75pRqannSSSe5eYcOHXK1bp769eu7/7dv3+6abIarTJkydvfdd7tJr1VCGPX/u+aaa1z/vUBdu3a15s2bu2AuVJAnhQoVchMAAAAA5NpkLKrBU1NL9a2TrVu32gMPPJBiGfXbU42fkrB4tXyPPvqo7dy507+M+vRdeOGFLmDzAkA19/zkk09cbV0o3377rY0dO9YFjXLgwAHXbFQ1fSVLlnTbDZwUwCnzpv4GAAAAgHgV80BPCVUUtHXp0sWqV69utWrV8idTCaRkLArgSpcubeXKlbOFCxe64K5AgQL+ZSZMmOBq8xo1auQSvCgoGzdunJ1++ukht63lfvzxR1erp+2qVvDyyy9P0VwTAAAAAHKaPD5lLIkD6g+n4Q4qV67s+uKtWrXKatasaYULF06xnPrIKTBT/7wqVaq45CcaFiF4XcrUqeeD+/WFcvjwYdu0aZNL8JJWs0v10VPzUtVChkt99FRDuHv3bitevHjYr0NKarqrRD0K9NVfE7HBeYgPu3btcjey2g6caPkLh+6DDOR20wczLFJ24bchPnAecsd5SIogtoh5Hz2PAqgTTjjB/zi4eeRXX33lAjyNjacD+Mgjj7gd9IZbCF6XaujCpVrBwKEaUqPAEwAAAADiXdwEeumpW7eua1KpWrW9e/e65psaV69ChQqxLhoAAAAAxJUcE+iphm7OnDkuAYsSpmj8OwAAAABADg70POqXAgAAAABIHVktAAAAACDBEOgBAAAAQILJcU03AQD/56P721vJkiU5HDFCKvP4wHkAgNCo0QMAAACABEOgBwAAAAAJhkAPAAAAABIMgR4AAAAAJBgCPQAAAABIMGTdBIAcqvuTMyx/4aKxLgZygemDO8a6CACACFGjBwAAAAAJhkAPAAAAABIMgR4AAAAAJBgCPQAAAABIMBkO9P7991975ZVXrH///v55P/30kx05ciRaZQMAAAAAZFegt3DhQjvxxBPtxRdftCeeeMI/f/z48fbWW2/F/ETs37/fVqxY4YLRtKxatcq2b9+e7vq2bt1qO3fujGIJAQAAACDOAr1+/frZ//73P/v1119TzL/lllts1KhRFms33nijC0Qfe+yxVJcZNGiQ1a9f35o0aWJr1qwJuYxqLOvWrWuNGjWyWrVquf+///77LCw5AAAAAMQo0FuwYIHdeuut7u88efL459epU8fVkmWlQ4cO2fr16+3w4cMhn3/jjTfs77//tpNOOinVdTzzzDM2evRo+/TTT+3000+39u3b2+bNm1Msc/ToUVu0aJFNmzbN1Qxu27bNLde5c2f3NwAAAAAkVKBXoEABS0pKOmb+8uXLrWzZspZVHnnkEStdurS1bt3aypUrZ/fff3+KPoHa/uDBg+2dd96xfPnyhVzH2LFjbdiwYfbVV19Zp06d7KOPPrIzzjjDzj///BTNM/X6l19+2WrXru0e58+f3/VH3L17t/3yyy9Zto8AAAAAEJNAr2PHjjZkyBBX6+XV6P3111+ulk81Xllh6dKl9vDDD9vs2bNt3bp1rt9c5cqVXeAlBw4csB49erg+gzVr1gy5jvnz59sLL7xg33zzjZ155pn+AG7cuHHWpk0b1yQ1OTk51TL8/vvv7v+qVatmyT4CAAAAQDTkz8iLnn76adeMsUKFCi4wUn+4P/74wxo3bmwjRoywrKAaRNWylSlTxl+reM899/ifv/fee61hw4Z2zTXXpLqOZs2auWanwfLmzeuacqZlz549duedd9pFF12UZrPQgwcPuimw3AAAAAAQ94Gemk2q+aL6uOl/BXunnnqqdenSxQVgWUHNK7t162b16tVzzSzbtm1rl19+uZUvX95mzJhh77//vn355Zcu26Yo2FJTTD1u0KBBprN4XnLJJS4gVLPQtCjQVc0jAAAAAOSoQE+1eQquunbt6qbsoCBrwoQJtnbtWvv6669d3zplzpw7d65LlqLaxeuvv96/vJbbsmWLe37ZsmWp9tlLj5qEqjmq1vXtt9/6axRTM2DAAFe7GFijV61atQxtGwAAAACyrY/enDlzbO/evZadvCyb6n/Xq1cv++KLL1ziF/1/9dVXu5q7wEm1eD179nR/ZzbI27hxo+vXp9rD9BQqVMiKFy+eYgIAAACAuA/0LrzwQvvggw8sO3322Weu+eTkyZNtyZIl9tprr7kArHnz5lmyPSWaUW2ltjVmzBh/M1BNXgIYAAAAAEiYppulSpWy3r1728cff+wSoBQsWDDF88OHD7doU9ClDJkaHkEDnFevXt0FfWeddVbI5TXAuZpzZpSSr2g72lcNBB+8f5deemmG1w0AAAAAcRfoaVB0jWX333//2bx58yy7XHzxxW4Kh4LAzChZsqQ/sQsAAAAAJHyg991330W/JAAAAACA2PXRAwAAAAAkWI3erbfemubzL7/8ckbLAwAAAACIRaC3bdu2FI81YPoff/xhS5cutQ4dOmS2TAAAAACA7A70Jk2aFHL+0KFDbdeuXZkpDwAAAAAgFoFeavr27Wsnn3yyjRo1KpqrBQCE8NH97V2GYMSGWrPs2LHDSpcubXnz0uUdABBfovrLtHXrVjf+HAAAAAAgh9Xohaqx27lzp7377rv00QMAAACAnBjovf7668fMK1WqlF1xxRU2YMCAaJQLAAAAAJCdgd6UKVOsTp06IZ9bvXq1FStWLKPlAQAAAADEoo9e3bp1M/QcAAAAACCHZd3cu3evFS1aNJqrBACkovuTMyx/Yb5zkbWmD+7IIQaARA/07rvvvpB/e2mmFy1aZE2bNo1e6QAAAAAAWRvoLV26NOTfUqBAATvppJPsnnvuibwUAAAAAIDYBHrTpk1z/19zzTVuKAUAAAAAQIIkYyHIAwAAAIAETMZy9OhRW7lypa1fv96OHDmS4rlOnTpFo2wAAAAAgOwK9DRWXrdu3WzZsmUuCUu+fPlc4CeFChWyAwcOZGS1AAAAAIBYNd28++67rWXLlrZnzx73WIHd3LlzrUmTJvbUU09ZIlm3bp2NHTvWnn/+efvmm2/M5/PFukgAAAAAEP1Ab/bs2fbII4/4x8xT8NO8eXMbN26cPfvss5Yohg0bZh06dLA5c+bYihUr7Nprr7Vzzz3X9u/fH+uiAQAAAEB0m27u2LHDKlSo4P4uW7asbdmyxapWrWp16tSxv//+O+L1ffDBB/bHH39Ynjx5rHz58nbWWWfZiSeeeMxyv/32m82YMcPKlStn5513ns2cOdNq1KhhZ599tn+ZnTt32qeffmrbtm2zevXq2UUXXeSalqbm8OHD9tlnn9maNWusVq1abvnChQu7584//3x78MEHXblk4MCBVq1aNbf+yy+/POL9BAAAAIC4rdEL1KxZM3viiSds7dq1NnLkSBfsRUrBlpp/qqbshx9+cLWDb7zxRopl3n//fTv11FNd7dr333/vmo4OGDDANaf06DkFdx999JELOB966CE788wzbd++fSG3e/DgQTvjjDNczd3GjRttwoQJbvn//vvPPd+iRQt/kCcFCxZ0j9UPEQAAAAASqkavV69e/r9HjBhhHTt2dH3YSpQoYRMnTox4fRqXL9CUKVPshhtusOuuu87y58/vgsB+/frZww8/7GrV5Ouvv3a1eh5l/uzRo4cL2m699VY3TwliFKypOemgQYOO2e78+fNdQpldu3b5a/FWrVqVIrj766+/7L333rOkpCSbPn2665/YuXPnVPdFwaMmj14HAAAAAHEf6L3++uv+v08++WQ3xIKmSpUqZbi2S/3+Fi9e7JqF7t271wVfWucJJ5xgv/76q/3zzz8pAsx27dpZzZo1UwRtSpyimjnVLKrfoCYFcKrpC6VixYouQBw/frxdddVVVqRIEVcjGEjBogJNBWwq2/bt210g5wWGwRT4KiAFAAAAgBzbdNOtJG9eF3RlNMi7+uqrrWvXri5YUybPQ4cOufkKrERBnrah/nvBgZpn06ZN/sBMTS8VLKrJppKnpFYDp2amqq0bM2aMlSlTxi0bPBh83bp1bfjw4fbyyy/bggULXP+8F154IdV9UXPS3bt3+6cNGzZk6JgAAAAAQLYPmD5p0iTXj05NGzVwulebddNNN7lkKeHSaxVsaWy+2rVru3nLly+3p59+OkVAp/H6lGAlcN1KAuPx5qu5Z/369cPevpKqaFISl48//thuvPFGK1WqlGuOGkyBZuPGjV3NY2oU7NKHDwAAAECOq9F78803XT84JWJRnzZPsWLFXLAXCdW8yXHHHeef98orr6RYRuPzKcjS8A2en376yWXK9KgvXuXKld32A8e60/p///33kNtWkKrgURTcqWmosmoq6FSCGPXfC7R582ZbsmSJNWrUKKJ9BAAAAIC4r9HToOiq0VNTRyU/8XTq1MllrXzmmWfCXpeCOAWMSqyiJpbqjxccYKnv3JNPPukCMT2vpC+ffPKJq+lTk05RLZoyc3bp0sVatWrlhlxQYKaAUOVp2LDhMdtWfzsNoaAgUX0B1XRUmT+7devmgkXV7qnfoV6rPoPK5tm0aVO78847M3LYAAAAACB+a/RUE6bgSAIzVKpWTMFTJJRVc9asWS6rphKcXHHFFa4vnLJkqobO07NnT5ewRc0yTzrpJDcMQ8mSJd3kUXD3559/2s033+xqCNu2besCvYsvvjjkthVgLlq0yDXT1LbVV1BNSVWrp6EUlMSlT58+LrBUsKemnRrO4fjjj8/AUQMAAACAOK7RUwCm5pCnnXZaikBv2rRp/n52kVCNnYZTCKQEKIE0Tp9q004//XT3WMGgmo2qBjGQgs3rr78+7G2ruakyboaifVONnyYAAAAASOhA75ZbbnHNGkePHu2CIdWCKcgbMmRIiqac0aSgTs0yNayCsmoqgYtq7rzADwAAAACQiUDvgQcecOPKdejQwQ1n0KBBA9dH7r777rM77rjDsoJq1apXr24zZsxw29Tg5cG1eQAAAACACAI9DSlwyimnuL9Vi/foo4+6fnRKnKKhD9SHTc0gs5ICSk0AAAAAgCgEeuof5w1boGBrxYoVVrRoUZfQBAAAAAAQP8LOuqnMk//884/72xsgHQAAAACQg2v0NMadxrzTeHNyxhlnpLqshiUAAGStj+5vn2KIGWQvdVvYsWOHlS5d2j+mKwAAOS7Qe/PNN23KlCm2evVqmzdvnhscHQAAAACQgwM9DWx+6aWX+mvsHnzwwawsFwAAAAAggzLU1kQ1ewAAAACA+ESnAgAAAABIMAR6AAAAAJBb++gBAOJL9ydnWP7CRWNdDCSg6YM7xroIAIBMokYPAAAAAHJrjd7dd98d9kpHjRqV0fIAAAAAALIr0NP4eQAAAACABAr0Pvvss6wtCQAAAAAgKuijBwAAAAAJJsNZN48ePWorV6609evX25EjR1I816lTp7DWMXz4cDv77LPdFG1PPvmknXLKKda+ffts3zYAAAAA5LhAT/31unXrZsuWLbPk5GTLly+fC/ykUKFCduDAgbDW8+6771rJkiWzJNj64IMPXJlSC/SyctsAAAAAkOOabioDZ8uWLW3Pnj3usQK7uXPnWpMmTeypp56yePDAAw/Y+eefH+tiAAAAAEDOqNGbPXu2/f7771a06P8N1Ovz+ax58+Y2btw4V9N3xx13RLS+L7/80q2zYMGCdv3111vVqlVTNLFs3bq1Cyrnz59v+fPnt549e1qNGjVs8uTJ7nXly5e3m266ydXQedatW2elSpXK9LZV47d3795UlwEAAACAhKjR27Fjh1WoUMH9XbZsWduyZYv7u06dOvb3339H3JfukUcescKFC9sPP/zg+tWtWbMmRRPLHj162IsvvugCy5kzZ9ppp51ml19+ub355ptWpkwZ+/jjj+3cc891zUgDm24uWLAg09u+9tpr01wGAAAAABIu62azZs3siSeesLVr19rIkSNdsBcJ9embNWuWDRw40KZNm2aNGjWyIUOGpFjmpJNOcjVv/fv3t08//dSSkpJs//797m/N++STT2zJkiU2b968qG87nGUCHTx40JUvcAIAAACAuA/0evXq5f97xIgRrkatVq1a9vzzz9uzzz4b0bq6d+/umkRKnjx57Morr7Rvv/02xTKBfe2KFy/uahMD55UrV84129ywYUPUtx3OMoF0PEqUKOGfqlWrFlGZAAAAACAmgd7rr7/u//vkk092QyyoOaOacEaaAEVNPwMpaPOagnqKFCmS4rGyfIaa52X+jOa2w1km0IABA2z37t3+KdLgEwAAAABiNo5eoLx581rNmjUz9NrgPn0KjLIr2Uk42460fGrqqQkAAAAA4j7Q05AKMmrUKP/fqdEy4VLSlMGDB1vp0qVdv7s33njDOnfubNkhnG3HsnwAAAAAkKWBngZJD/V3Zql2TFk0NYSCxuLTUA2DBg2K2vozu+1Ylg8AAAAAsjTQ++yzz0L+nRmqKdMg62rqqDHyLr74YuvYsaN/fL7AZQJpuIPTTz89xbzHH3/cBWSBA6anlQE0nG2LxuxTv8O0lgEAAACAhOujl1FXX321/+969eqlu0xg8JVWJlDROHuZ3Xbg8+ktAwAAAAA5PtBbtWqVzZ4923bu3HnMc+n14QMAAAAAxFmgN2bMGLvjjjusUqVKbvy6RA30QjUbBQAAAICEDPSGDRtm7733nl1xxRWWyEI1GwUAAACAhBwwXQOBd+rUKfqlAQAAAADEJtBr166dTZ8+PfNbBwAAAADER9PNF154wVq0aGFTp0612rVrW548eVI8/+CDD0arfACAVHx0f/uQ/aSRPZKTk23Hjh1WunRpy5s3Q/dNAQCIr0Dvueees82bN9ucOXNsxYoVxzxPoAcAAAAAOSzQe+ONN1xtngYQBwAAAADElwy1NcmfP7+1bds2+qUBAAAAAMQm0GvZsqWr0QMAAAAAJEjTTXU879mzp02ZMsXq1KlzTDKW4cOHR6t8AAAAAIDsCPTWrl1rrVq1si1btrgJAJD9uj85w/IXLsqhR6ZMH9yRIwgACShDgd53330X/ZIAAAAAAKKCgX8AAAAAIMFkqEZP1q9f78bTW758ufl8PmvYsKHdddddVr169eiWEAAAAACQ9TV6arpZv359mz59ulWqVMmqVKni/tY8mnUCAAAAQA6s0XvggQfsvvvus2HDhqWYP3jwYPfcvHnzolU+AAAAAEB21OgtWbLE7r333mPma96vv/5qsbZ//37buXNnVNe5a9cu27t3b1TXCQAAAABxE+iVLFnS/vzzz2Pm//HHH+65WPn555+tbdu2VrlyZatVq5b7/9VXX83UOn/88Uc79dRTrWrVqnbCCSfY9ddfb3v27IlamQEAAAAgLgK9q6++2i6//HKbMGGC/fXXX256//333byrrrrKslpycrL9+++/7v9A3377rT322GO2fft2VwP3xBNP2C233OICwIysTzWX7du3ty5durgaws2bN7vHCmgBAAAAIF5lKNAbMWKEC36uu+46q127tptU09WtWzcbOXKkZaXHH3/c1Ro2btzYypQp4/oFHj161D03aNAgO+OMMyxv3v/brSuvvNLy589vK1asyND6HnzwQVeb99BDD1mBAgUsT548LsjVPAAAAABIqECvUKFC9swzz9iOHTts8eLFruZLf2tewYIFLatoKIcBAwbYN998Y1u2bLGNGzda0aJFU/THU/+8tWvXur6Cd999t2ty2blz54jXd/jwYZs5c6Z1797d1fTpeS8ABAAAAICEHEdPjjvuODv55JMtuyiYVM2aN1afgjIFaoGU8VM1jWq+qdq8sWPHWtmyZSNe3z///GMHDhywrVu3ur55+/bts6SkJNc09fnnn3f7HsrBgwfd5NFrAAAAACDuAz3VbL377rv2008/uWAp2KRJkywrtGzZ0i666CKrV6+edejQwSVeUXPR0qVL+5c555xzXI2ejB8/3i677DL79NNP7cILL4xofWqmKW+++aZ9//33boxAJaA5++yzbeDAgTZ69OhUm7U+/PDDWbL/AAAAAJBlTTfvueceu/POO2337t2utix4yirqezdlyhSbM2eOtWrVyt577z2rW7duqslR1J+uefPmNnHixIjXV65cOddEVf38FOSJ+iJee+21Nm3atFTLqBpBHRdv2rBhQ5T2HgAAAACysEZPGTbVf01BVHZSTWK+fPmsQYMGbrr99ttds0rV2Cn4FK8mTnw+n2t6mVrylLTWpzEBVXunYC2QsnkWK1Ys1TIqONQEAAAAADkq0FNNWMOGDS27qfZNtXPK8KmATP3xNm3aZKeddppLwtKuXTvr16+fK5sCspdeesklWLn55psjXp8MGTLELrjgAjvzzDNdjd/cuXPt7bffdn30AAAAACChAr3zzjvPPvnkE9c0Mjt5GTAVaK1Zs8YlUdFYfuqXJ6+88oo99dRTro+ckqU0bdrUZQVVc8yMrE/BnWr3NGTE008/7Z5XoKfxAgEAAAAgoQK9448/3tWCTZ061erUqZOiuaQMHz7csoqSq2gKpUmTJjZu3LiorU/atGnjJgAAAABI6EBv5cqVrrZLY8tpAgAAAADk8EDvu+++i35JAAAAAACxG15BfdQ0xID6twEAAAAAEiDQ07AEnTt3tho1atiDDz7oBhIHAAAAAOTgQO+jjz5ywxBorDklZFFWSyUseeedd9wwBwAAAACA2Mnj06jimTR//nx788033VS4cGG78sorrW/fvnbiiSdabpeUlGQlSpRwA68XL1481sXJsdRMeMeOHVa6dGk3jiM4D7mZxgktVaqU7dy500qWLBnr4uRafC/FB85D7HEO4gPnIXech6QIYotMb11ZN2fNmuUmDbNw0UUX2apVq+ykk06y0aNHZ3b1AAAAAIAIZSjQO3LkiGuy2aVLF6tataq99dZbdvPNN9vGjRvdgOPffPONS9YyZMiQjKweAAAAAJDdwysouPvvv//siiuusO+//95atmx5zDLt27d3zTgBAAAAADkg0HvkkUdcP7xixYqludzmzZszWi4AAAAAQHYGemqmCQAAAADI4YHe3XffHfZKR40aldHyALnL0K6W0zr1lo11IWB5Dx7mKAAAgOgEeqtXrw53UQAAAABATgj0Pvvss6wtCQAAAAAgKhh5GgAAAAASTIaSscj69evtueees+XLl5vP57OGDRvaXXfdZdWrV49uCQEAAAAAWV+j991331n9+vVt+vTpVqlSJatSpYr7W/P0HAAAAAAgh9XoPfDAA3bffffZsGHDUswfPHiwe27evHkRrW/JkiVWsWJFq1ChgkXb0qVLrXTp0la5cuWorxsAAAAAEqZGT4HZvffee8x8zfv1118jXt8VV1xhH374oWWF66+/3saNG5epdfz77782Z84cS0pKilq5AAAAACCuAr2SJUvan3/+ecz8P/74wz0XTxo3buyalmbEokWLrEePHtaoUSNr2bKlLVy4MOrlAwAAAIC4CPSuvvpqu/zyy23ChAn2119/uen9999386666qoMF+bo0aO2atUqW7t2bchaxC1btlhycrILKNesWeN/7uDBg64mcfPmzce8rl+/ftauXbs0t7tz5073+uAaOwV6l1xyic2dOzfD+wQAAAAAOSLQGzFihHXp0sWuu+46q127tpvURLJbt242cuTIDBVESVyUsVPrPfHEE926FMAFNu+86aab3DJ6Tlk+u3btalOmTHHbV/BZs2ZNGzhwYERNN9XctFq1anbttdfaCSecYL1797bDhw+752688Ua78sorrWDBghnaJwAAAADIEclY/v77b/v222/tlFNOcbVgBw4csDx58rhg67jjjstwQZS186effrImTZrYunXrXFPJ559/3iV98SxevNj1latatar98ssv1qxZM1eG+fPnu+yfs2bNsjZt2riAUEFberQ+DRGhWkRv+bFjx9qePXtcApeMUHAaGKDSrw8AAABAXNfo/fjjj64mrWfPnq42T4GWgiIFZ5kJ8rwaO61HatSoYbfeequ9+eabx9TOKciT008/3cqXL2+9evVyQZ6cc845VqRIEVu2bFlY2zx06JALUtUc1HPDDTdkOMjzajtLlCjhn1RbCAAAAABxG+g9+OCDrmmlslBq6ty5sxtSIRo0Bl+gBg0auL5/gYKHX1BQp2AveN7evXvD2mbz5s1doKiELW3btrXhw4eH7B8YiQEDBtju3bv904YNGzK1PgAAAADI0kBPCVGefvppK1eunJv0t+ZFQ3BwpsfFixe3rPbyyy+75qi33367qwlU/8DM7FOhQoVcuQMnAAAAAIjbQG/Xrl0uwAusYVPGymiYOXNmisdfffWVnXbaaZaV9u3b5/4vU6aMde/e3WUOVdPQ4LIAAAAAQEInY1E/vfTmnXXWWREXZMGCBXbLLbe4gEvJXiZNmuSSq2Slr7/+2p566inX90/JWObNm+dq91q3bu2e37p1qxsvUP/L77//boULF3bBoNdXEAAAAAByfKDnBUFpzfP5fBGtUxk81f9PffKeffZZN5zBl19+aWeeeWaKZSpWrJjidU2bNj2mj56StJQtWzasAdMvvvhiN8C7Mm1qCAYN3TBjxgzXd08U+A0bNsz93aJFC7eMJmX11AQAAAAA8SiPL4KobPXq1WEtV6dOncyUKaFoeAVl31RiFvrrZZwyo+7YscNlRM2bN0PDP8anoV1jXQLkQEkHD1uJkZ+7pvO6WYXYSNjvpRyG8xB7nIP4wHnIHechKYLYIqIaPQI4AAAAAIh/3IIEAAAAgARDoAcAAAAACYZADwAAAAASDIEeAAAAAOT24RUARNHQyTnqcJLRKz4k79plNrJUrIsBAADiGDV6AAAAAJBgCPQAAAAAIMEQ6AEAAABAgiHQAwAAAIAEQ6AHAAAAAAmGrJtAtAztmivuDJWNdSFgeQ8e5igAAIA0UaMHAAAAAAmGQA8AAAAAEgyBHgAAAAAkGAI9AAAAAEgwBHoAAAAAkGByTKD3xBNP2E8//RTrYgAAAABA3Msxwyu8+eabVrRoUWvVqlW2b/vnn3+2b775xm3/iiuusCpVqmR7GQAAAAAg4Wr0YiE5OdluuOEG69y5s+3cudNNHTp0sBUrVsS6aAAAAACQ82v0PKpZmzNnjhUsWNCuueYaq1ixYormnarxO3jwYKrLBNu7d69NnDjR1qxZY7Vq1bJLL73UihUr5p576aWX3HOLFy+2unXrunn333+/7d+/Pxv2FAAAAAByQY3eqFGjXKB14MAB++yzz+zkk0+29evXp2jeed1116W5TKB9+/bZ6aefbmPHjrUCBQq44FCB4p49e9zzL7zwgl111VX+IE+KFy9uFSpUyIa9BQAAAIBcUqOnhCyFCxd2zSpbt25tQ4cOdQFepMvIggULXE3er7/+6gI9+fvvvy1fvnwu2Fu5cqUNHDjQpkyZ4patXLmydevWLc1AT7WJmjxJSUlR3X8AAAAASKgaPTWrVAAnefPmtauvvtpmzpwZ8TKeatWq2dGjR+25556z7du3u3lVq1Z1SVd2797tHj/99NM2ZswYFwiqhlC1e/PmzUu1jCNGjLASJUr4J20DAAAAALJTjgr0ypUrl+KxatY2b94c8TKemjVr2qeffmpffvmlC8hOOeUU11xTvH56pUuXtunTp9tDDz1kn3/+uWvaOWDAgFTLqOcUJHrThg0bMry/AAAAAJDwTTc3bdqU4rGaWQYPdRDOMoEuvPBCN6m5pWrslLzFa6Kp/5s1a5ZieT1+5513Ul1foUKF3AQAAAAAsZKjavQ++OADf5NKBWZKotKxY8eIl/GoD55X46bgrGvXrq4GcOPGjW5ejx49bNasWa6vn/h8Pvvuu++sSZMmWbqfAAAAAJBravTULFNZMtu0aWOzZ892QyMMHjw44mU8GiZBY+TVqVPHTjjhBPvll19cIpbLL7/cPa/mmu3atbPTTjvNWrZs6frmbd261QWPAAAAABCvckyg179/f2vatKnlz5/f5s6d67JpdunSxd+XztOrVy8799xz01zGoz55yrj59ddfu+yb559/vl1wwQVu/D1RMhUFi9OmTbN169a5Jp5axkv2AgAAAADxKMcEejfccIP/75NOOinNZfV8est41GSzQ4cOqT6vbJsXX3xxBCUFAAAAgNjKUX30AAAAAAAJVKMXSfNOAAAAAMjNEirQC2zeCQAAAAC5FU03AQAAACDBEOgBAAAAQIJJqKabQEwNnZzwJyA5Odl27NhhpUuXtrx5uU8Us/Owa5fZyFIx2z4AAIh/XKkBAAAAQIIh0AMAAACABEOgBwAAAAAJhkAPAAAAABIMgR4AAAAAJBiybiJxDO0a6xLkijtDZWNdCFjeg4c5CgAAIE3U6AEAAABAgiHQAwAAAIAEQ6AHAAAAAAmGQA8AAAAAEgyBHgAAAAAkmIQM9A4dOmSbN292/0fLkSNH3Dr37NkTtXUCAAAAQFZIqEBv3bp19r///c+qV69ulSpVsp9//jlq67711lvdOgcPHhy1dQIAAABAVsixgV5SUtIx8yZOnGglSpSwL774IirrC1zv4sWLrWHDhhGvFwAAAACyW44L9EaPHm1lypSxatWqWbly5ezRRx+15ORk99z9999vAwYMsPLly0dlfbJmzRq7++67bfz48VagQIEs2ScAAAAAyLWB3sqVK+2ee+6xqVOn2u7du2316tWuH9727duzZH2HDx+2Hj162JAhQ6x+/fpR3hsAAAAAyBr5LQfZunWrq1U78cQT3WM103z44YezbH0DBw60ChUq2C233BL2Og8ePOimcJqEAgAAAIDl9hq9li1bWtu2ba1evXp244032rvvvpupQCqt9c2aNcveeOMN15RT2TY1KfPmvn373N+pGTFihAsYvUlNQgEAAAAgO+WoQC9fvnz25Zdf2syZM61Ro0Y2ZswYq1u3rv31119RX5+mggULWvv27e2UU05xk5p6qq+e/j569GjIdaqPoJqBetOGDRsyudcAAAAAkMCBnpckRYFWv3797Mcff7SiRYvalClTor6+G264wV+T500KBnv37u3+VpAYSqFChax48eIpJgAAAADITjmqj97kyZNd4pTrr7/eTjjhBJs3b54Luk4++WT3/P79+10tmvreyY4dO9zzxx9/vJsiXR8AAAAA5EQ5KtDr2rWr6yM3bNgwN+yBBkYfO3astWvXzj2voK1v377ubyVR6dOnj/v7vvvuc1Ok6wtWtmxZaugAAAAAxL08Pp/PF+tCJDIld1FSFtU00owz49TMVjW0pUuXtrx5U2lxPLRrJrYA5BxJBw9biZGf286dO61kyZKxLk6uFdb3EjgPuQCfhfjAecgd5yEpgtiCXyYAAAAASDAEegAAAACQYAj0AAAAACDBEOgBAAAAQIIh0AMAAACABEOgBwAAAAAJJkeNowekaehkDlAWI3VzfEjetctsZKlYFwMAAMQxavQAAAAAIMEQ6AEAAABAgiHQAwAAAIAEQ6AHAAAAAAmGZCxIHEO7xroEueLOUNlYFwKW9+BhjgIAAEgTNXoAAAAAkGAI9AAAAAAgwRDoAQAAAECCIdADAAAAgARDoAcAAAAACYZADwAAAAASDMMrpGPNmjU2Z84cy58/v7Vo0cKqV6+ePWcGAAAAADKIGr1U+Hw+u+yyy+z888+3Tz/91MaPH2/169e3p556KqPHGgAAAAByT43e559/7mrO8uTJY+XLl7czzjjDqlWrdsxya9eutW+++cbKlStnZ511ln333XdWpUoVa968uX+Zffv22cyZM23btm1Wr149t1x6AZ3Wo+3XqlXLzj77bMuXL5+bf8UVV9jEiRNdueS9996za665xrp27Wq1a9fOgiMBAAAAAAlSo7dhwwZbsWKFLV++3N555x1r2LChTZgwIcUyU6dOdTVqmj9u3Dg7/fTT7Z577rEvvvjCv8ySJUtccPf000/bjz/+aDfeeKO1b9/eDh06FHK7hw8ftnPPPdduueUW++mnn2zkyJHWunVrFyzmzZvXLr30Un+QJ+edd54LAFeuXJmFRwMAAAAAEqBG79Zbb03xWMHcnXfe6ZpOqnZNgZoe33ffffboo4/6A79LLrnE/5rk5GS7/PLLXfDXr18/N0+vU0D43HPPudcGmz9/vut/t337djv++OPdvIULF9rRo0dDllPbVHlOPvnkVPfl4MGDbvIkJSVFfDwAAAAAIMcHevL777+7GrkdO3bYnj17XNNL1fTVrFnTfv31V1u/fr3dfvvt/uU7d+6connnL7/8YqtWrXJB2ssvv+xq3jSpmadq90IFeqVLl7YjR47Y9OnTrUuXLi6IO/XUU0OWb9myZS6AvP/++11z0dSMGDHCHn744UwfDwAAAADI0U03FcC1atXKPvzwQ/vtt9/szz//dPNV0yYbN250TSkrVaqU4nWBAZeCQlm3bp0tXbrUBWYKHk888UQ755xzQm63QYMGLijs37+/lS1b1gV7n3322THLrV692iVlUXDp1SimZsCAAbZ7927/5JULAAAAAHJNjZ6CupdeeskFZ40aNXLz1Afu9ddfdzVyogQtapq5c+dOVwvnUa2fx5t/7733RpQopXfv3m5SgDhp0iSXaOXLL790/fG88qkfn6a33nrLBZxpKVSokJsAAAAAINfW6O3atcv9X7FiRf88JVsJ1LhxYxfIffDBB/556kunmjaPMnWqmeYzzzyT4rXqp/fXX3+F3Pbff/9t//33n/u7Ro0armmmmooq6BS9TgGeagRVJjXtBAAAAIB4F/MavSZNmriavA4dOlj37t1dfzwNoRBIiVKGDRtmd999t8vOWaJECZedU80tvayYRYoUccGY1qEATcHZ5s2b7auvvrJHHnnETjjhhGO2rVq8Nm3auGaZen7evHmuj6CaaO7fv989p0BRwzeMGTPG/7q2bdu6zKAAAAAAEI9iXqNXoEABN7RBjx49XGDWokULl1hF/fbUZNPTp08f16RSzSIV+OlvPa+gz3PhhRfaH3/8YRdccIFt3brVBW8afkHDJISifoE///yzCzS1bY25pyEe9Do1Fb344otd5k+tUwGmN6nvHQAAAADEqzw+ryNcnNu0aZNLxuLV4CnRimoDf/jhB2vZsqXFKw2voGBUwWHx4sVjXZwcS4G3alvVhDfVfpJDu2Z3sYCYSDp42EqM/Nz1Wy5ZsiRnIZ6/l8B5yAX4LMQHzkPuOA9JEcQWMW+6Ga4FCxa4Ac1Va6d+dW+88YYbNy+egzwAAAAAiIUcE+ipGaUStkybNs3fH0/9+gAAAAAAOTTQk2bNmrkJAAAAAJA6OhUAAAAAQIIh0AMAAACABJOjmm4CaRo6mQOUxcjoFR+Sd+0yG1kq1sUAAABxjBo9AAAAAEgwBHoAAAAAkGAI9AAAAAAgwRDoAQAAAECCIdADAAAAgARD1s3cbGhXy0l3JMrGuhDgPMSJvAcPx7oIAAAgzlGjBwAAAAAJhkAPAAAAABIMgR4AAAAAJBgCPQAAAABIMAR6AAAAAJBgCPQAAAAAIMEwvEIakpOTbcKECTZ9+nT777//rFGjRnbrrbda5cqVs+8MAQAAAECEqNFLw3XXXWfffPONXXjhhdajRw+bO3eunXbaabZp06ZIjzMAAAAAZJtcX6O3atUqe/HFF23NmjVWq1Yt69Onj9WvX98dnBdeeMFKlCjhP1gXX3yxFStWzL766iu7/vrrs+8sAQAAAEAEcnWN3o4dO+zMM8+0w4cPW+/eva1BgwZ21VVXufkSGOTJr7/+akeOHLETTzwxRiUGAAAAgPTl6ho9BW7qe/fSSy/55914442WN+//H//+8ssv9r///c+SkpJcrd+kSZOsRYsWqa7z4MGDbvLodQAAAACQnXJ1jZ6SqxQuXNh69uzp+uLt37/fChUqZAUKFPAvo+acCvTuvfdea9q0qQ0YMCDNPnojRoxwNYHeVK1atWzaGwAAAAD4P7k60CtXrpwtWLDAypcvb/369bOyZcvaTTfdZIcOHfIvU6ZMGTvvvPNcMpbPP//cNd0cNWpUqutUILh7927/tGHDhmzaGwAAAAD4P7m66abUrl3bnnrqKff36tWrrVmzZtayZUvr1avXMcuqpk9DK2zevDnV9alGUBMAAAAAxEqurtGbM2eOzZgxw/+4YsWKLkjz+Xx24MABe+WVV9xYeh4175w/f76r4QMAAACAeJWra/QqVapkQ4YMcUMlnHDCCbZ8+XJXm3f11Ve72js9rlKliuunt2vXLtu4caMNGjTIrr322lgXHQAAAABSlasDvRo1atj06dNdchVl1FTilOrVq/ufV1+8YcOG2e+//27HHXeca+ZZpEiRmJYZAAAAANKTqwM9j/rdaQpFA6SnNZwCAAAAAMSbXN1HDwAAAAASEYEeAAAAACQYAj0AAAAASDAEegAAAACQYAj0AAAAACDBkHUzNxs62XIKDVy/Y8cOK126tOXNy/0JzkPulrxrl9nIUrEuBgAAiGNcMQMAAABAgiHQAwAAAIAEQ6AHAAAAAAmGQA8AAAAAEgyBHgAAAAAkGLJu5kZDu1pOvCNRNtaFAOchTuQ9eDjWRQAAAHGOGj0AAAAASDAEegAAAACQYAj0AAAAACDBEOgBAAAAQIIh0AMAAACABJOQWTePHj1qe/bsseOPP97y58/4Lu7fv98OHjyYYl6+fPmsWLFiUSglAAAAAGSNhKrR27Rpkw0dOtRq1KhhpUqVsh9//DFT67vnnnusfPnyVrNmTf90wQUXRK28AAAAAJAVcmygd+jQoWPmvf322+bz+WzSpElRWZ907tzZdu3a5Z9+/vnnDJUXAAAAALJLjgv0Xn75ZatSpYoVL17c/f/0009bcnKye27AgAH28MMPW9WqVaOyPo+abyqABAAAAICcIEcFen/88Yf16dPHxo0bZwcOHLBFixbZv//+a9u3b8+y9X366acuCFR/v/PPP99+//33KO4RAAAAAOTyQG/z5s1WoEABO/30091j9Z97/PHHrVy5clmyvsaNG9vMmTNt3759tnr1apeEpW3btrZt27ZU16nav6SkpBQTAAAAAGSnHBXonXHGGdaqVStr0KCBq4lTX7y9e/dm2fpuv/12a926tcu0WalSJVfzp8Bt4sSJqa5zxIgRVqJECf9UrVq1DJcPAAAAABI+0FPt29dff22TJ092/elGjhxp9evXt3Xr1mXL+o477jjX/2/NmjWprlP9BHfv3u2fNmzYkKGyAQAAAECuCPQkT548riZu0KBBNm/ePDdO3scff5wt69u6dasLAqtXr57q+goVKuT69AVOAAAAAJCdclSg99FHH9ktt9xis2fPti1bttjnn3/u/m/UqJF/iAQNgeD1i/vvv//cYyVaiXR96mun/nhfffWVbdy40S3TtWtX13/vmmuuydb9BgAAAIBI5LccpEuXLrZjxw7r16+faz6pmjUNj6BsmKI+duprJ+of5wVk//vf/9wU6fqGDRvmkrMoG6cGYFd/PQWH+hsAAAAA4lUeHwPEZSnVLiroVH+9uGnGObRrrEsAIBOSDh62EiM/t507d1rJkiU5ljGiMVd1s7B06dKWN2+OaiCTUDgPscc5iA+ch9xxHpIiiC34ZQIAAACABEOgBwAAAAAJhkAPAAAAABIMgR4AAAAAJBgCPQAAAABIMAR6AAAAAJBgctQ4eoiSoZNz3KEkZXB84DzEh+Rdu8xGMp4nAABIHTV6AAAAAJBgCPQAAAAAIMEQ6AEAAABAgiHQAwAAAIAEQ6AHAAAAAAmGQA8AAAAAEgyBHgAAAAAkGAI9AAAAAEgwBHoAAAAAkGAI9AAAAAAgwRDoAQAAAECCIdADAAAAgARDoAcAAAAACYZADwAAAAASDIEeAAAAACQYAj0AAAAASDAEegAAAACQYAj0AAAAACDBEOgBAAAAQILJH+sCJDqfz+f+T0pKinVRcrTk5GTbs2eP5c+f3/Lm5f4E5yF3875P9D+fh9jheyk+cB5ij3MQHzgPueM8JP2/awAvxkgLgV4W04mWatWqZfWmAOQyNWrUiHURAABAjGKMEiVKpLlMHl844SAyFdVv2rTJihUrZnny5OFIZuLuhYLlDRs2WPHixTmOMcJ5iA+ch/jAeYgPnIfY4xzEB85D7jgPPp/PBXmVK1dOt8aQGr0sphNQtWrVrN5MrqEPDIFe7HEe4gPnIT5wHuID5yH2OAfxgfOQ+OehRDo1eR46OwEAAABAgiHQAwAAAIAEQ6CHHKFQoUI2ZMgQ9z84D7kdn4f4wHmID5yH2OMcxAfOQ3woFEfXrCRjAQAAAIAEQ40eAAAAACQYAj0AAAAASDAEegAAAACQYBhHDznGwYMHbeHChVa+fHmrXbt2yGU0gOTKlSutXLlyVqNGjWwvY24zb948O3ToUIp51atXdxOyho73smXLrEiRItagQQMOczb777//bPHixcfMP/nkk61YsWKcjyy2fPly27Fjh7Vq1SrVgYS1jD4nJ510kuXPz2VOVli0aJH7v2nTpsc89+OPPx4zr06dOlaxYsUsKUtutXXrVvv777+tVq1aVrJkyZDL7N+/337//Xc35prOAaJP50DfSbouPe6441I8p/k6/sGaN29uBQsWtOxAMhbEvV27dtnIkSPt3XffdX9fc8019vLLLx+z3CuvvGL33nuvCzLWr19vbdu2tQ8++MCKFi0ak3LnBvrh1g+IAmvPddddZ717945puRLV119/bVdeeaUdf/zxlpSUZFWrVrVPP/3UqlWrFuui5Rq//PKLNWvWzM444wzLly+ff/6rr75qDRs2jGnZEpm+y5999llbtWqVe+8fOXLkmGVWr15tl1xyibsA1o2Qw4cP24cffphqUIjIjRo1yv3+btmyxapUqWJLly5N8bzOS4ECBVyQHTig8/333+/ODTJv7ty59sADD7gbGjoHurmt3wWdFx17z6RJk+ymm25yN8d1vpo0aWKffPKJlS5dmtMQBfrtHThwoO3cudMd07/++sv69etnDz/8sH+ZKVOmWPfu3a1ly5YpXjt58uQU101ZygfEuaVLl/pGjBjh27Jli69Vq1a+W2655ZhlFi1a5MuTJ49v4sSJ7rGWrV69uu+ee+6JQYlzjwoVKvjeeeedWBcjV9i5c6evVKlSvoEDB7rHBw8e9LVu3drXpk2bWBctV5k/f75PP506H8g+Q4YM8f3888++sWPH+vLlyxdymWbNmvk6dOjgO3LkiHt8++23+ypWrOjbu3cvpypK+vbt6/v99999/fv39zVq1OiY5w8fPuw+H99++y3HPIvoN/f777/3P/7jjz98ZcqU8T388MP+eevWrfMVKlTI99xzz7nHSUlJvpNOOsl35ZVXcl6iZPTo0e761DNr1ixfgQIFfB988IF/3uTJk33HHXecL5boo4e416hRI/vf//7n7kqlZuzYsVa/fn277LLL3GMte/PNN9tbb71lycnJ2Vja3Ed3z1XL8e+//8a6KAlNdwD37t1r/fv3d4/V7EN/f/vtt7Z27dpYFy/XUe2RmnDqnCDrDR069Ji74oF+++03mz9/vrvD7tW0Dho0yH0vffHFF5yiKNbonXjiiekut2nTJluwYIFruoboUqum1q1b+x+rSWb79u1TNJkdP368a/nRp08f91jNyu+++25Xy6cacWTeXXfd5a5PPWeffbZ7HKrpsmpf9R114MABy24EekiY/gKnnXbaMW2gVaW+bt26mJUrt1yAqXmI+gmouSzHO+ve43Xr1rXixYuneI97zyF7devWzTWXKlWqlLuACtWUENnH+wwE/g5UqlTJNW/m8xGbi+AbbrjBnQM1Xdu+fXsMSpE7HD161L3HA/vg6fEpp5ySonm5fi/UnFl9vBF9ut78888/j+kLqZuBF198sXXt2tX9XgwbNsyyE72Uke3Uzy64XX8wJVKJpN+R7hoGdwovU6aM/zkFIUjfnDlz0rxgVUfjwOP82GOPWc+ePV3CA9Xs6cusR48e9vPPP1uePHk45FGk97H3nvZ4fS24a5599EM9c+ZMa9eunT8hkf7WuRk8eHA2lgSB9BlQf+zChQunmK/zwucj++h7/4033nBBnv7WjT/VNqmFzUcffcSbNgsMGTLEJQS555570vy9CLwmQnQpCZTe4/p9UJ4Cj65j1eLJuwH15ZdfWufOnd0NKH1GsgOBHrKdOg6rKWZa9IFRABEudUAOrhJXtinJrsxGiUCdiJW5NDUnnHCCjRs3zv/4xhtv9P+tjsXDhw93P+pqSkhwHV2h3uPeY97j2UeZ1QKz/uouea9evWzChAkEejH+fCgzsy64Am8y6XeAz0f2UQ1S4O+CbtqqOa1afej7KjgQR+a88MIL9uSTT9rHH3/sWnx4uCbK/hrsb775xk2BSYiCW5pddNFFLimRfi8I9JCwWrRoEbINc2box2Tjxo0p5nmPyUgYPt1tyowKFSr4jz2BnkX9PT5jxoyQ73GGs4gtve+Dv3+Q/Z8PNWFTdkEvjb/6Z2/evJnPRxx8PnRudC5q1qwZ6+IkjDFjxrgsj8os27Fjx2M+D6pJCsTvRdZQTar6RKqlh4bZCefzsGLFCssu9NFDQlAt0nfffZeiNkpphJUGPbXxZZA5oZJQfPXVV64ZJ+O7Zc17XD/UGksy8D2uPnu6eYLYve8VgCudPGJHySkKFSpkU6dO9c+bNWuW6yqgzw5i+7ug32E1V0N0aDgp9Q2eOHGiawoYTO959dNTk87A3wvdFKxXrx6nIUo0pNfbb7/tfgNOPfXUdD8P6hqja9Xs/L2g6Sbinpri/PTTT+5vZYvSXUHVCKo/hvfBUhX4888/76rE+/bt6+5kadylzNZQIXX6slIGtquuusoqV67szpGakAwYMMDKli3LoYsyjQWmPpBKAKLO3Nu2bXN9Mx599FGaQ2Ujvb/1Y61+ebqpofE9Z8+ebdOmTcvOYuQ6Gj9PGTT/+OMP99hrFaKxwXSzQ82l1ERQ47Wp6aYyDqqLgL6ftAyiY8mSJe6GqgKIffv2+c+DxpXU50E1G9OnT3eJJ9QnTH+/+OKLrokhg9dHh75zbrvtNrvvvvvcMfbOQWAf+i5durhm5fpfnwtlCX7uuedcUEL/+eh46KGHbPTo0W58TzUR986Drn+8m91qxqza1bPOOsslwtFYh8pIqwA9uzBgOuLeoUOHXDbHYPrw6EfFo2QgGlhdKc/VX0xfhOecc042lzZ3UdIVDW2hAep1p1AXVW3atIl1sRKW+rjoR0VDKmhAaCW+UeCH7KMmaO+88467iaS7tfpBv/POO933EbKObm4oaAjVfK1x48b+xxpSR0k/9Ltx/vnnu3NDH73oUV+7UM3OPv/8c3/fJH02dKNVN2XVr1t9WIP7KiH6n4XgPvS6Ma6br7oRpXOjG+KdOnXi0GfxZ0HXQF5mTfUbfu2111zfPTUlV02eKiOybbB0Aj0AAAAASDz00QMAAACABEOgBwAAAAAJhkAPAAAAABIMgR4AAAAAJBgCPQAAAABIMAR6AAAAAJBgCPQAAAAAIMEQ6AEAEMDn87lBhidOnGi//fabm7d79243KLQGgt6+fbt999139uuvv4Z93NasWWNTp07Nscc5p5cfAHIjAj0AQI7w119/2WeffWbTp0+3tWvXZtl2LrvsMrv22mvto48+suXLl7vArkGDBjZixAibPHmy7dy504YPH27vvfde2OucNWuW9enTJ0uOyaeffmpZLavKDwDIOvmzcN0AAEQlmLnhhhts8eLF1rJlS8uXL5+tXLnSqlWrZq+//rrVrl07akd5x44dLsBTbV3jxo3dvHHjxlnBggXtxx9/9C/Xpk0bO+GEE8Jeb61ateySSy6xaPvmm29c0HnxxRdHfd0AgJyNQA8AELe2bdtm55xzjjVt2tQ2bNhgxYsX9z/3ww8/2J49e1IsrwBt9erVVqlSJWvRooXlzXtsw5V169a5oLFMmTJ26qmnWtGiRf3zFeR56162bJn7e8aMGZYnTx6bMGGCfx0KLhs2bHjMulXGRYsWWYUKFez00093QalUr17dLrjggrDLIn/88Yebzj//fLdf//zzj51yyilWpUoV9/zGjRtt/vz5tnfvXn/ZTj75ZDvxxBNTBK5fffWVde7cOcW6VUup/VKAqGOopqhSpEgRq1+/vqvBTMvSpUtty5Yt1q5dO/+89evX24IFC6xr165h7yMAIOsQ6AEA4taoUaNs165druYuMMiT1q1b+/8+cuSIXX755S5gOeOMM1xgVLFiRfvyyy+tXLly/r53d999t7377rtuGQWRCp4U3DVr1swFaV9//bVbVs1DFfSIag/VR2/KlCn+7Wk7119/vTVp0sS/7r59+7pyKsA8cOCA5c+f3/XrU7nV9PHBBx90AVc4ZfHK8Oijj1rNmjWtUKFCbt7cuXNdP0GtR8srqNy3b5+/bAqiAgO94447zm677TZ3fK655hr//DfffNNefPFFu+KKK1xg7L1eQeNPP/1kF110kSubAtxQJk2aZDNnzkwR6P388892xx13+AO9cPYRAJCFfAAAxKnTTz/dd+GFF6a73AsvvOArXbq0b926de5xUlKSr0mTJr7evXv7l3n11Vd9tWvX9m3bts0/b/jw4b4GDRr4Hy9fvtynn8YNGzb45z355JO+k08+OcX22rVr5+vfv7//8YsvvugrWrSob8mSJf55P//8s+/ff/91f48dO9ZXpUqViMry/PPPu7JMmjTJP+++++5z++V57bXXfDVq1Ejz2PTq1euYY6j9GTBgQMjlt27d6qtUqVKK7QaXf8iQIb5WrVqleN3777/vK1OmTET7CADIOtToAQDi1ubNm11zv/So6aJqrNREUooVK2Z33XWX3Xvvvfbqq6+6eWPHjnU1cN9++62rbdJ0/PHH24oVK1xNk5p7ZtTbb79tPXv29NfwifoTpibcsqg2snv37v7XnXvuufb8889HVDYdl/POO8/+/fdfK1++vGuSumTJkhTJZA4fPuyaXW7atMkOHTrk+j/OmzcvxbYjlZXHGwCQPgI9AEDcUsCm/mTpUT+wSy+9NMU89aNLSkpyWTJLlSrlMnUqiFGzw0Bqvqj5maH+aT169Ah7+XDLUrp06RTPqwnnwYMHIyqb+jhWrlzZNfm88847bfz48a7Po9fHUENIdOjQwd8/T8091f9OgWFmZOXxBgCkj0APABC3mjdv7vrDqY+Z+rylpmzZsi7xSCA9LlCggL9vn/5v27atPfHEE1EvZ8mSJcMKSD1ZWZZg6md35ZVXugBPfehUk6faTs+gQYNcWVQr6bnwwgtdDVxqlOQmOTk5xTz1S4zVPgIAjsU4egCAuKWARNkln3322WOeU83W1q1b3d9nnXWWSygSGHx8+OGH/uEYvOBFwU5wpk6tP7OUGVM1ZoE1VUpsokQpoUSrLGoKGRxgpdZ8U4lcFMz9/fffLvALbB6rmrzAx0rIkhZl/tQg6kePHvXPUxPNQFl5vAEA6aNGDwAQt9Q/Txkib775Zps9e7YbokC1SatWrXKDl+s59WNTRksFdmqC2K1bN7fsJ5984h82QB566CGXVVMZH3v37u0yVCr40Th933//fabKOXjwYJs2bZoLLK+77joXfL3//vv2xRdfhBxOIFpl0fFRsKux9OrUqXPM8AoejQmoSU03VcsW2D+uS5cu9vTTT1vhwoVdran6AKZVeyrK+tmvXz+76qqr3DlRYKj9z4p9BABkDIEeACCuXXvttS44UcIVJQhRc0zVQGkAcw2hIEoyorHaXnnlFf/8hQsXphgPTv3d9Hql+1fAocBGSUoC+9apuaH6kAUGZ1qHaqcCBQ+Yrqajv/zyiws8NeSB+sSpb5oXUAUPmB5OWerVq2cdO3ZMsV3tl8oXuIyCyalTp7okK8HDKwQaOHCgq/XUsBCB/ve//1nVqlXdEBDqAzh69GhXWxc4BmFw+RVcaww/JbrRfqhGVYPaa3D5SPYRAJB18ij1ZhauHwAAAACQzeijBwAAAAAJhkAPAAAAABIMgR4AAAAAJBgCPQAAAABIMAR6AAAAAJBgCPQAAAAAIMEQ6AEAAABAgiHQAwAAAIAEQ6AHAAAAAAmGQA8AAAAAEgyBHgAAAAAkGAI9AAAAALDE8v8BRle+Es7oHQcAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                                                        model  L2_norm  max_abs_coef  n_zero  n_coefs\n",
      "                                                          OLS  1311.15        811.11       0       65\n",
      "                  <class 'sklearn.linear_model._ridge.Ridge'>    56.25         24.38       0       65\n",
      "     <class 'sklearn.linear_model._coordinate_descent.Lasso'>    42.48         23.21      29       65\n",
      "<class 'sklearn.linear_model._coordinate_descent.ElasticNet'>    29.38         15.94      15       65\n"
     ]
    }
   ],
   "source": [
    "from matplotlib.patches import Patch\n",
    "\n",
    "coef_summary = []\n",
    "\n",
    "for name, pipe in fitted_models.items():\n",
    "    # TODO 1: names of the polynomial features (get_feature_names_out of the \"poly\" step with the\n",
    "    #         input names X_train.columns) and the coefficients of the \"reg\" step (coef_)\n",
    "    feature_names = pipe.named_steps[\"poly\"].get_feature_names_out(X_train.columns)\n",
    "    coefs = pipe.named_steps[\"reg\"].coef_\n",
    "\n",
    "    coef_table = pd.Series(coefs, index=feature_names)\n",
    "\n",
    "    # TODO 2: number of coefficients that are zero within the tolerance (np.isclose with atol=1e-10, rtol=0)\n",
    "    n_zero = int(np.isclose(coefs, 0, atol=1e-10, rtol=0).sum())\n",
    "\n",
    "    # TODO 3: the 15 coefficients with the largest absolute value, sorted by their value (for the bar chart)\n",
    "    top = coef_table.loc[coef_table.abs().nlargest(15).index].sort_values()\n",
    "\n",
    "    coef_summary.append({\n",
    "        \"model\": name,\n",
    "        \"L2_norm\": np.linalg.norm(coefs),\n",
    "        \"max_abs_coef\": np.abs(coefs).max(),\n",
    "        \"n_zero\": n_zero,\n",
    "        \"n_coefs\": len(coefs),\n",
    "    })\n",
    "\n",
    "    # use different colors for positive and negative coefficients\n",
    "    colors = np.where(top >= 0, \"steelblue\", \"coral\")\n",
    "\n",
    "    fig, ax = plt.subplots(figsize=(9, 6))\n",
    "    ax.barh(top.index, top.values, color=colors)\n",
    "    ax.axvline(0, color=\"black\", linewidth=0.8)\n",
    "\n",
    "    ax.set_xlabel(\"Coefficient value\")\n",
    "    ax.set_ylabel(\"Polynomial feature\")\n",
    "    ax.set_title(f\"{name}: 15 largest coefficients by absolute value\\n\"\n",
    "                 f\"Zero coefficients: {n_zero} of {len(coefs)}\")\n",
    "\n",
    "    ax.legend(handles=[Patch(facecolor=\"steelblue\", label=\"Positive coefficient\"),\n",
    "                       Patch(facecolor=\"coral\", label=\"Negative coefficient\")])\n",
    "\n",
    "    ax.grid(axis=\"x\", alpha=0.25)\n",
    "    ax.set_axisbelow(True)\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.show()\n",
    "\n",
    "# summary table for all models\n",
    "print(pd.DataFrame(coef_summary).round(2).to_string(index=False))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0c2a7b06",
   "metadata": {},
   "source": [
    "**Output 2.2**\n",
    "\n",
    "Neregularizovaný model (OLS) má extrémne veľké koeficienty (L2 norma 1311, najväčší 811) –\n",
    "model na seba „naťahuje“ šum a navzájom sa vyrovnávajúce kladné a záporné váhy, čo je typický\n",
    "prejav pretrénovania.\n",
    "\n",
    "Všetky tri regularizačné metódy koeficienty výrazne zmenšili (L2 norma 29–56, max 16–24).\n",
    "- **Ridge** koeficienty zmenšil, ale žiadny nevynuloval (0 z 65).\n",
    "- **Lasso** vynulovalo 29 z 65 koeficientov → vznikol riedky model, Lasso zároveň vybralo príznaky.\n",
    "- **Elastic Net** vynuloval 15 z 65 – menej než Lasso, lebo jeho penalizácia je len z polovice L1.\n",
    "\n",
    "Nulovanie treba posudzovať podľa stĺpca `n_zero`, nie podľa grafov – grafy ukazujú\n",
    "len 15 najväčších koeficientov, nulové sa v nich neobjavia.\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "39449e4d",
   "metadata": {},
   "source": [
    "## Krok 3: Regularizačná cesta a stabilita výberu korelovaných príznakov\n",
    "V Kroku 2 sme porovnávali modely pri pevne stanovenej hodnote hyperparametra alpha. V tejto časti sa zameriame na dva aspekty metódy Lasso, ktoré nemožno dostatočne posúdiť na základe jediného nastavenia regularizácie. Najskôr preskúmame priebeh odhadovaných koeficientov v závislosti od hodnoty hyperparametra alpha, označovaný ako regularizačná cesta. Následne vyhodnotíme stabilitu výberu príznakov pri miernych zmenách trénovacej vzorky v prítomnosti silnej korelácie medzi príznakmi."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4783ecde",
   "metadata": {},
   "source": [
    "### Úloha 3.1 - Regularizačná cesta modelu Lasso\n",
    "Preskúmajte, ako hodnota hyperparametra alpha ovplyvňuje koeficienty modelu Lasso. Pre zachovanie čitateľnosti grafu použijeme 10 pôvodných príznakov bez polynomického rozšírenia. Funkcia lasso_path vypočíta koeficienty pre postupnosť hodnôt alpha, čo umožňuje sledovať priebeh koeficientu každého príznaku pozdĺž regularizačnej cesty.\n",
    "\n",
    "Pracujte iba s trénovacou množinou X_train, y_train. Vstupné príznaky štandardizujte a cieľovú premennú centrujte odčítaním jej priemeru na trénovacej množine, pretože funkcia lasso_path neodhaduje absolútny člen.\n",
    "\n",
    "Doplňte časti kódu označené komentárom TODO v kostre nižšie. Kód následne vypíše poradie prvého vstupu príznakov do modelu na použitej mriežke hodnôt alpha. Za vstup považujeme situáciu, keď absolútna hodnota koeficientu pri znižovaní alpha prvýkrát prekročí toleranciu 1e-10. Viaceré príznaky môžu vstúpiť v rovnakom kroku.\n",
    "\n",
    "V grafe identifikujte hodnoty alpha, pri ktorých sú všetky koeficienty numericky nulové, a opíšte, ako sa koeficienty menia pri postupnom oslabovaní regularizácie. Poradie vstupu príznakov interpretujte ako vlastnosť konkrétnej regularizačnej cesty, nie ako všeobecné poradie ich dôležitosti.\n",
    "\n",
    "Poradie vstupu príznakov do modelu závisí aj od ich vzájomných korelácií, preto ho nemožno interpretovať ako jednoznačné poradie ich dôležitosti. Regularizačná cesta slúži na analýzu priebehu koeficientov v závislosti od sily regularizácie. Vhodnú hodnotu hyperparametra alpha následne zvolíme na základe validačnej chyby odhadnutej pomocou krížovej validácie.\n",
    "\n",
    "Upozornenie\n",
    "V tomto grafe hodnota alpha smerom doprava klesá, a teda regularizácia slabne. Smerom doľava sa regularizácia zosilňuje a podporuje zmenšovanie koeficientov až na nulu. Pri interpretácii preto sledujte hodnoty na osi, nielen smer pohybu v grafe."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "94cf237a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Order in which the features enter the model (alpha at which the coefficient becomes non-zero):\n",
      "bmi    44.306\n",
      "s5     35.112\n",
      "bp     22.051\n",
      "s3     15.557\n",
      "sex     6.136\n",
      "s6      3.854\n",
      "s2      3.854\n",
      "s1      3.430\n",
      "age     1.520\n",
      "s4      1.353\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.linear_model import lasso_path\n",
    "\n",
    "# lasso_path fits no intercept, so the features are standardized and the target is centered\n",
    "# TODO 1: standardize the original training features (StandardScaler, fit_transform on X_train)\n",
    "#         and center the target (y_train minus its mean, as a NumPy array)\n",
    "X_train_scaled = StandardScaler().fit_transform(X_train)\n",
    "y_train_centered = (y_train - y_train.mean()).to_numpy()\n",
    "\n",
    "# TODO 2: coefficient paths with lasso_path for alphas=np.logspace(2, -3, 100) and max_iter=50_000;\n",
    "#         it returns (alphas, coefs, dual_gaps), where coefs has the shape (n_features, n_alphas)\n",
    "alphas_path, coefs_path, _ = lasso_path(\n",
    "    X_train_scaled, y_train_centered, alphas=np.logspace(2, -3, 100), max_iter=50_000\n",
    ")\n",
    "\n",
    "# entry of each feature: the largest alpha with |coefficient| > 1e-10\n",
    "# (alphas_path is decreasing, so it is the first such position on the path)\n",
    "entered = np.abs(coefs_path) > 1e-10\n",
    "entry_alpha = pd.Series(\n",
    "    [alphas_path[row.argmax()] if row.any() else np.nan for row in entered],\n",
    "    index=X_train.columns,\n",
    ").sort_values(ascending=False)\n",
    "print(\n",
    "    \"Order in which the features enter the model (alpha at which the coefficient becomes non-zero):\"\n",
    ")\n",
    "print(entry_alpha.round(3).to_string())\n",
    "\n",
    "# plot one regularization path per original feature\n",
    "fig, ax = plt.subplots(figsize=(10, 6))\n",
    "\n",
    "for feature, path in zip(X_train.columns, coefs_path):\n",
    "    ax.plot(alphas_path, path, label=feature)\n",
    "\n",
    "# moving right corresponds to weaker regularization\n",
    "ax.set_xscale(\"log\")\n",
    "ax.invert_xaxis()\n",
    "ax.axhline(0, color=\"grey\", linestyle=\"--\", linewidth=0.8)\n",
    "\n",
    "ax.set_xlabel(\"alpha (right = weaker regularization)\")\n",
    "ax.set_ylabel(\"Coefficient value\")\n",
    "ax.set_title(\"Lasso regularization path\")\n",
    "ax.legend(bbox_to_anchor=(1.02, 1), loc=\"upper left\")\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8d23b12f",
   "metadata": {},
   "source": [
    "**Output 3.1**\n",
    "\n",
    "Pri `alpha` väčšom ako približne 44 sú všetky koeficienty numericky nulové – regularizácia\n",
    "je taká silná, že model nepoužije žiadny príznak a predikuje iba priemer.\n",
    "\n",
    "Pri postupnom znižovaní `alpha` vstupujú príznaky do modelu postupne: prvý `bmi` (alpha ≈ 44),\n",
    "potom `s5` (≈ 35), `bp` (≈ 22), `s3` (≈ 16), `sex` (≈ 6), `s6` a `s2` súčasne (≈ 3.9),\n",
    "`s1` (≈ 3.4) a nakoniec `age` (≈ 1.5) a `s4` (≈ 1.4).\n",
    "Po vstupe sa absolútne hodnoty koeficientov pri ďalšom oslabovaní regularizácie zväčšujú,\n",
    "niektoré však nemenia veľkosť monotónne.\n",
    "\n",
    "Poradie vstupu je vlastnosť tejto konkrétnej regularizačnej cesty – závisí aj od korelácií\n",
    "medzi príznakmi (napr. s1, s2, s4 sú krvné hodnoty, ktoré spolu súvisia), preto ho nemožno\n",
    "chápať ako jednoznačné poradie dôležitosti príznakov.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5fc7449d",
   "metadata": {},
   "source": [
    "### Úloha 3.2 - Stabilita výberu pri korelovaných príznakoch\n",
    "V predchádzajúcej úlohe sme sledovali zmeny koeficientov pri rôznych hodnotách hyperparametra alpha. Teraz ponecháme jeho hodnotu konštantnú a preskúmame, ako výber príznakov ovplyvňujú zmeny trénovacej vzorky.\n",
    "\n",
    "K pôvodným príznakom v množine X_train pridajte dve takmer identické kópie príznaku bmi. Každú vytvorte pridaním nezávislého náhodného šumu s malou smerodajnou odchýlkou vzhľadom na variabilitu bmi. Kópie vytvorte iba raz, pred bootstrapovaním. Polynomické rozšírenie v tejto úlohe nepoužívajte.\n",
    "\n",
    "Následne vytvorte 50 bootstrapových vzoriek. Každú získajte náhodným výberom rovnakého počtu pozorovaní, aký obsahuje trénovacia množina, s opakovaním. Vstupy a zodpovedajúce cieľové hodnoty vyberajte spoločne pomocou rovnakých indexov.\n",
    "\n",
    "Na každej vzorke natrénujte Pipeline s krokmi StandardScaler a Lasso, pričom vo všetkých opakovaniach použite rovnakú hodnotu alpha. Zaznamenajte, ktoré z troch príznakov - pôvodný bmi a jeho dve kópie - majú absolútnu hodnotu koeficientu väčšiu než 1e-10.\n",
    "\n",
    "Doplňte časti kódu označené komentárom TODO v kostre nižšie."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "0b3c03d6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "            selection_count  selection_frequency\n",
      "bmi                      32                 0.64\n",
      "bmi_copy_2               29                 0.58\n",
      "bmi_copy_1                1                 0.02\n"
     ]
    }
   ],
   "source": [
    "from sklearn.utils import resample\n",
    "\n",
    "# create two strongly correlated copies of BMI\n",
    "rng = np.random.default_rng(RANDOM_STATE)\n",
    "X_corr = X_train.copy()\n",
    "# noise SD = 5 % of the bmi SD -> the copies correlate with bmi at about 0.999\n",
    "noise_std = 0.05 * X_train[\"bmi\"].std(ddof=0)\n",
    "\n",
    "# TODO 1: add the columns \"bmi_copy_1\" and \"bmi_copy_2\" to X_corr: the bmi column plus independent\n",
    "#         noise rng.normal(0, noise_std, len(X_corr)) for each copy\n",
    "X_corr[\"bmi_copy_1\"] = X_corr[\"bmi\"] + rng.normal(0, noise_std, len(X_corr))\n",
    "X_corr[\"bmi_copy_2\"] = X_corr[\"bmi\"] + rng.normal(0, noise_std, len(X_corr))\n",
    "\n",
    "# track how often each feature has a non-zero coefficient\n",
    "tracked = [\"bmi\", \"bmi_copy_1\", \"bmi_copy_2\"]\n",
    "selection_counts = pd.Series(0, index=tracked, dtype=int)\n",
    "n_bootstrap = 50\n",
    "\n",
    "for seed in range(n_bootstrap):\n",
    "    # TODO 2: bootstrap sample - rows of X_corr and y_train drawn with replacement\n",
    "    #         (resample with replace=True and random_state=seed keeps features and targets aligned)\n",
    "    X_boot, y_boot = resample(X_corr, y_train, replace=True, random_state=seed)\n",
    "\n",
    "    # TODO 3: pipeline with the steps \"scaler\" (StandardScaler) and \"lasso\" (Lasso(alpha=2.0, max_iter=50_000)),\n",
    "    #         fitted on the bootstrap sample\n",
    "    model = Pipeline(\n",
    "        [(\"scaler\", StandardScaler()), (\"lasso\", Lasso(alpha=2.0, max_iter=50_000))]\n",
    "    ).fit(X_boot, y_boot)\n",
    "\n",
    "    coefs = pd.Series(model.named_steps[\"lasso\"].coef_, index=X_corr.columns)\n",
    "    # TODO 4: boolean mask of the tracked features with a non-zero coefficient\n",
    "    #         (negated np.isclose with atol=1e-10, rtol=0)\n",
    "    selected = ~np.isclose(coefs[tracked], 0, atol=1e-10, rtol=0)\n",
    "    selection_counts += selected.astype(int)\n",
    "\n",
    "# report selection counts and frequencies across bootstrap samples\n",
    "selection_results = pd.DataFrame(\n",
    "    {\n",
    "        \"selection_count\": selection_counts,\n",
    "        \"selection_frequency\": selection_counts / n_bootstrap,\n",
    "    }\n",
    ")\n",
    "\n",
    "print(selection_results.sort_values(\"selection_count\", ascending=False))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aad4c8c2",
   "metadata": {},
   "source": [
    "**Záver 3.2:**\n",
    "\n",
    "\n",
    "Pôvodný príznak `bmi` a jeho dve takmer identické kópie (korelácia ≈ 0.999) nesú rovnakú\n",
    "informáciu. V 50 bootstrapových vzorkách Lasso vybralo `bmi` 32-krát, `bmi_copy_2` 29-krát\n",
    "a `bmi_copy_1` len 1-krát.\n",
    "\n",
    "Výber medzi silne korelovanými príznakmi je teda nestabilný – závisí od konkrétnej trénovacej\n",
    "vzorky, nie od skutočnej dôležitosti príznaku. Nulový koeficient v Lasso preto nedokazuje,\n",
    "že príznak je nedôležitý; jeho informáciu mohol prevziať korelovaný „dvojník“.\n",
    "Pri korelovaných vstupoch je vhodnejší Elastic Net (má tendenciu ponechať skupinu spolu)\n",
    "alebo Ridge (rozdelí váhu medzi všetky korelované príznaky).\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8b82a093",
   "metadata": {},
   "source": [
    "## Krok 4: Výber sily regularizácie pomocou krížovej validácie\n",
    "Doteraz sme skúmali, ako regularizácia ovplyvňuje koeficienty modelu a stabilitu výberu príznakov. Teraz sa zameriame na výber jej sily. Príliš slabá regularizácia nemusí dostatočne obmedziť pretrénovanie, zatiaľ čo príliš silná môže viesť k podtrénovaniu.\n",
    "\n",
    "Vhodnú hodnotu hyperparametra alpha budeme hľadať pomocou krížovej validácie na trénovacích dátach. Najskôr zobrazíme závislosť trénovacej a validačnej chyby od hodnoty hyperparametra alpha pomocou validačnej krivky. Následne vyberieme konfiguráciu podľa validačnej RMSE. Odloženú testovaciu množinu použijeme až na záverečné vyhodnotenie zvoleného modelu.\n",
    "\n",
    "Ladenie ukážeme na modeli Ridge s polynomickými príznakmi druhého stupňa. Tento výber slúži didaktickému cieľu a neznamená, že daná konfigurácia prekonáva všetky modely z predchádzajúcich úloh."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4a6efbd4",
   "metadata": {},
   "source": [
    "### Úloha 4.1 - Validačná krivka\n",
    "Pre model Ridge s polynomickými príznakmi druhého stupňa preskúmajte hodnoty hyperparametra alpha z mriežky alpha_grid = np.logspace(-4, 4, 30). Použite Pipeline s krokmi PolynomialFeatures, StandardScaler a Ridge, pričom nastavte degree=2 a include_bias=False.\n",
    "\n",
    "Pracujte s pôvodnými príznakmi z množiny X_train, bez pridaných kópií príznaku bmi z Úlohy 3.2. Pomocou funkcie validation_curve vyhodnoťte všetky nastavenia na rovnakých rozdeleniach päťnásobnej krížovej validácie ako v predchádzajúcich porovnaniach. Použite scoring=\"neg_root_mean_squared_error\" a pri spracovaní výsledkov zmeňte znamienko skóre, aby ste získali hodnoty metriky RMSE.\n",
    "\n",
    "Doplňte časti kódu označené komentárom TODO v kostre nižšie.\n",
    "\n",
    "Kód vypíše hodnotu alpha s najnižšou priemernou validačnou hodnotou metriky RMSE. Túto hodnotu best_alpha použijeme v Kroku 5 na záverečné vyhodnotenie.\n",
    "\n",
    "\n",
    "Upozornenie\n",
    "Pásma okolo kriviek vyjadrujú variabilitu medzi foldmi, nie intervaly spoľahlivosti.\n",
    "\n",
    "Nesprávny postup je vypočítať testovaciu hodnotu metriky RMSE pre každé alpha a ponechať najlepšie. Hoci sa na teste nefitujú koeficienty, informácia z testu by riadila výber modelu a výsledná testová chyba by už nebola nezávislým odhadom.\n",
    "\n",
    "Tip\n",
    "V praxi sa na výber hyperparametra často používa GridSearchCV. Robí to isté ako výber minima validačnej krivky (rovnaká mriežka, rovnaké foldy, rovnaká metrika) a pri refit=True navyše natrénuje model so zvolenou hodnotou na celej trénovacej množine. My tento krok urobíme ručne v Úlohe 5.1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "11554c52",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Best alpha from the grid: 62.102\n",
      "Mean validation RMSE: 56.94\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.model_selection import validation_curve\n",
    "\n",
    "# TODO 1: pipeline with the steps \"poly\" (degree=2, include_bias=False), \"scaler\" and \"ridge\" (Ridge());\n",
    "#         the step name \"ridge\" determines the parameter name \"ridge__alpha\" below\n",
    "ridge_pipe = Pipeline(\n",
    "    [\n",
    "        (\"poly\", PolynomialFeatures(degree=2, include_bias=False)),\n",
    "        (\"scaler\", StandardScaler()),\n",
    "        (\"ridge\", Ridge()),\n",
    "    ]\n",
    ")\n",
    "\n",
    "\n",
    "# evaluate regularization strengths on the same CV splits\n",
    "alpha_grid = np.logspace(-4, 4, 30)\n",
    "\n",
    "# TODO 2: validation_curve for the parameter \"ridge__alpha\" over alpha_grid,\n",
    "#         on the same folds cv and with scoring=\"neg_root_mean_squared_error\"\n",
    "train_scores, val_scores = validation_curve(\n",
    "    ridge_pipe,\n",
    "    X_train,\n",
    "    y_train,\n",
    "    param_name=\"ridge__alpha\",\n",
    "    param_range=alpha_grid,\n",
    "    cv=cv,\n",
    "    scoring=\"neg_root_mean_squared_error\",\n",
    ")\n",
    "\n",
    "# convert negative scores to RMSE\n",
    "train_rmse = -train_scores\n",
    "val_rmse = -val_scores\n",
    "\n",
    "# summarize results across folds for each alpha\n",
    "train_mean = train_rmse.mean(axis=1)\n",
    "train_sd = train_rmse.std(axis=1, ddof=1)\n",
    "val_mean = val_rmse.mean(axis=1)\n",
    "val_sd = val_rmse.std(axis=1, ddof=1)\n",
    "\n",
    "# TODO 3: index of the alpha with the lowest mean validation RMSE and the corresponding alpha\n",
    "best_idx = np.argmin(val_mean)\n",
    "best_alpha = alpha_grid[best_idx]\n",
    "\n",
    "print(f\"Best alpha from the grid: {best_alpha:.5g}\")\n",
    "print(f\"Mean validation RMSE: {val_mean[best_idx]:.2f}\")\n",
    "\n",
    "# plot mean RMSE and fold variability\n",
    "fig, ax = plt.subplots(figsize=(9, 5))\n",
    "ax.plot(alpha_grid, train_mean, color=\"steelblue\", label=\"Training RMSE\")\n",
    "ax.fill_between(\n",
    "    alpha_grid,\n",
    "    train_mean - train_sd,\n",
    "    train_mean + train_sd,\n",
    "    color=\"steelblue\",\n",
    "    alpha=0.2,\n",
    ")\n",
    "ax.plot(alpha_grid, val_mean, color=\"coral\", label=\"Validation RMSE\")\n",
    "ax.fill_between(\n",
    "    alpha_grid, val_mean - val_sd, val_mean + val_sd, color=\"coral\", alpha=0.2\n",
    ")\n",
    "\n",
    "# mark the best value from the evaluated grid\n",
    "ax.axvline(\n",
    "    best_alpha, color=\"grey\", linestyle=\"--\", label=f\"Selected alpha = {best_alpha:.3g}\"\n",
    ")\n",
    "ax.scatter(best_alpha, val_mean[best_idx], color=\"coral\", edgecolor=\"black\", zorder=3)\n",
    "\n",
    "# moving right corresponds to stronger regularization\n",
    "ax.set_xscale(\"log\")\n",
    "ax.set_xlabel(\"alpha (right = stronger regularization)\")\n",
    "ax.set_ylabel(\"RMSE\")\n",
    "ax.set_title(\"Validation curve of Ridge regression\")\n",
    "ax.legend()\n",
    "ax.grid(alpha=0.25)\n",
    "\n",
    "plt.tight_layout()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "64f1b98b",
   "metadata": {},
   "source": [
    "**Záver 4.1:**\n",
    "\n",
    "Validačná krivka má typický tvar písmena U:\n",
    "- pri malých `alpha` (slabá regularizácia) je trénovacia chyba nízka, ale validačná vyššia\n",
    "  (~62.5) – model je pretrénovaný, správa sa ako OLS,\n",
    "- pri veľkých `alpha` (silná regularizácia) rastú obe chyby (validačná ~74.8) – model je\n",
    "  podtrénovaný, koeficienty sú stlačené takmer na nulu,\n",
    "- najnižšia priemerná validačná RMSE 56.94 je pri `alpha` ≈ 62.1.\n",
    "\n",
    "So silnejúcou regularizáciou sa generalizačná medzera (rozdiel medzi krivkami) zmenšuje.\n",
    "Hodnota `alpha` bola vybraná iba pomocou krížovej validácie na trénovacích dátach –\n",
    "testovacia množina zatiaľ nebola použitá, aby jej výsledok zostal nezávislým odhadom.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f94b7109",
   "metadata": {},
   "source": [
    "## Krok 5: Záverečné vyhodnotenie modelu\n",
    "Po dokončení ladenia vyhodnotíme zvolený model na dátach, ktoré neboli použité pri učení ani pri výbere hyperparametrov. Odložená testovacia množina poskytne záverečný odhad jeho generalizačnej schopnosti."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ebdae892",
   "metadata": {},
   "source": [
    "### Úloha 5.1 - Jednorazové vyhodnotenie na testovacej množine\n",
    "Model Ridge s hodnotou best_alpha z Úlohy 4.1 natrénujte na celej trénovacej množine a jediný raz ho vyhodnoťte na odloženej testovacej množine. Vypočítajte:\n",
    "\n",
    "testovaciu hodnotu metriky RMSE,\n",
    "testovaciu hodnotu metriky MAE,\n",
    "koeficient determinácie \n",
    ".\n",
    "Doplňte časti kódu označené komentárom TODO v kostre nižšie.\n",
    "\n",
    "(Prečo sa testová RMSE líši od priemernej CV RMSE a aké obmedzenia má záver založený na jednej malej dátovej sade a jednom testovacom rozdelení, prediskutujeme na cvičení.)\n",
    "\n",
    "Výsledok testu je konečný\n",
    "Po zobrazení testových výsledkov už podľa nich neupravujte model, hodnotu alpha ani rozdelenie dát. Aj nepriaznivý výsledok je súčasťou korektného vyhodnotenia.\n",
    "\n",
    "Dokumentácia: clone, root_mean_squared_error, mean_absolute_error, r2_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "e9966b04",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Selected alpha:       62.1\n",
      "Mean CV RMSE (4.1):   56.94\n",
      "\n",
      "Test RMSE:  51.12\n",
      "Test MAE:   40.75\n",
      "Test R^2:   0.507\n",
      "Difference test - CV RMSE: -5.83\n",
      "RMSE of the reference model (mean of y_train): 73.22\n"
     ]
    }
   ],
   "source": [
    "from sklearn.base import clone\n",
    "from sklearn.metrics import mean_absolute_error, r2_score, root_mean_squared_error\n",
    "\n",
    "# final model: a copy of the Task 4.1 pipeline with the selected alpha, fitted on the whole training set\n",
    "# (clone keeps ridge_pipe itself unchanged)\n",
    "final_model = clone(ridge_pipe).set_params(ridge__alpha=best_alpha)\n",
    "final_model.fit(X_train, y_train)\n",
    "cv_rmse = val_mean[best_idx]\n",
    "\n",
    "# record the selected configuration BEFORE looking at the test set\n",
    "print(f\"Selected alpha:       {best_alpha:.4g}\")\n",
    "print(f\"Mean CV RMSE (4.1):   {cv_rmse:.2f}\")\n",
    "\n",
    "# TODO 1: predictions of final_model for X_test\n",
    "y_pred = final_model.predict(X_test)\n",
    "\n",
    "# TODO 2: test RMSE, MAE and R^2 (compare y_test with y_pred)\n",
    "test_rmse = root_mean_squared_error(y_test, y_pred)\n",
    "test_mae = mean_absolute_error(y_test, y_pred)\n",
    "test_r2 = r2_score(y_test, y_pred)\n",
    "\n",
    "print(f\"\\nTest RMSE:  {test_rmse:.2f}\")\n",
    "print(f\"Test MAE:   {test_mae:.2f}\")\n",
    "print(f\"Test R^2:   {test_r2:.3f}\")\n",
    "print(f\"Difference test - CV RMSE: {test_rmse - cv_rmse:+.2f}\")\n",
    "\n",
    "# reference model: always predicts the training mean\n",
    "baseline_rmse = root_mean_squared_error(y_test, np.full(len(y_test), y_train.mean()))\n",
    "print(f\"RMSE of the reference model (mean of y_train): {baseline_rmse:.2f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "756d68e0",
   "metadata": {},
   "source": [
    "**Záver 5.1:**\n",
    "\n",
    "Model Ridge s polynomickými príznakmi 2. stupňa a `alpha` ≈ 62.1 (zvolené krížovou validáciou\n",
    "v úlohe 4.1) dosiahol na odloženej testovacej množine:\n",
    "\n",
    "| metrika | hodnota |\n",
    "|---|---|\n",
    "| RMSE | 51.12 |\n",
    "| MAE | 40.75 |\n",
    "| R² | 0.507 |\n",
    "\n",
    "Referenčný model, ktorý vždy predikuje priemer trénovacích hodnôt, má RMSE 73.22 – zvolený\n",
    "model je teda výrazne lepší než triviálny odhad a vysvetľuje približne polovicu variability\n",
    "cieľovej premennej.\n",
    "\n",
    "Testová RMSE (51.12) je nižšia než priemerná CV RMSE (56.94). Rozdiel je pravdepodobne\n",
    "spôsobený malou testovacou množinou (89 pozorovaní) a konkrétnym rozdelením dát –\n",
    "jeden test na malej vzorke je zaťažený veľkou náhodnou variabilitou.\n",
    "Testovacia množina bola použitá iba raz a podľa jej výsledku sa model ďalej neupravuje.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7225fcb0",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "pok-isi (3.14.5)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.14.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}
