{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "c000",
   "metadata": {},
   "source": [
    "# Logistic Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c001",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:47.804890Z",
     "iopub.status.busy": "2026-11-17T23:43:47.804788Z",
     "iopub.status.idle": "2026-11-17T23:43:49.054305Z",
     "shell.execute_reply": "2026-11-17T23:43:49.053845Z"
    },
    "tags": [
     "hide-input",
     "thebe-init",
     "runnable"
    ]
   },
   "outputs": [],
   "source": [
    "path_data = '../../data/'\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "plt.style.use('fivethirtyeight')\n",
    "%matplotlib inline\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.pipeline import make_pipeline\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix, precision_score, recall_score\n",
    "\n",
    "# The shopping session data and the train-test split from Section 1501.\n",
    "shoppers = pd.read_csv(path_data + 'online_shoppers.csv')\n",
    "shoppers['purchase'] = (shoppers['outcome'] == 'purchase').astype(int)\n",
    "features = ['page_value', 'exit_rate', 'product_pages']\n",
    "X = shoppers[features]\n",
    "y = shoppers['purchase']\n",
    "X_train, X_test, y_train, y_test = train_test_split(\n",
    "    X, y, test_size=0.3, random_state=14, stratify=y)\n",
    "\n",
    "# Error costs from Section 1502.\n",
    "COST_FN = 30     # dollars lost for each buyer the model misses (an assumption)\n",
    "COST_FP = 3      # dollars wasted for each non-buyer the model flags (an assumption)\n",
    "\n",
    "def error_cost(actual, predicted):\n",
    "    \"\"\"Total dollar cost of the false negatives and false positives.\"\"\"\n",
    "    tn, fp, fn, tp = confusion_matrix(actual, predicted).ravel()\n",
    "    return COST_FN * fn + COST_FP * fp\n",
    "\n",
    "# The fitted models and their test-set probabilities, so the exercises also work\n",
    "# when the cells above them have not been run. The cells below refit the same models.\n",
    "logit = LogisticRegression().fit(X_train, y_train)\n",
    "p_buy = logit.predict_proba(X_test)[:, 1]\n",
    "knn = make_pipeline(StandardScaler(), KNeighborsClassifier(n_neighbors=11)).fit(X_train, y_train)\n",
    "p_buy_knn = knn.predict_proba(X_test)[:, 1]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c003",
   "metadata": {},
   "source": [
    "A nearest neighbor classifier answers \"purchase\" or \"no purchase\". For many business decisions, a **probability** is more useful: a session with a 90% chance of ending in a purchase deserves a nudge right now, one with a 15% chance might get a cheaper one, and one with a 1% chance can be left alone. Section {doc}`1502-evaluating-classifiers` also suggested that when a missed buyer costs much more than a false alarm, it may pay to flag sessions whose chance of a purchase is below 50%. That requires a classifier that reports how likely each outcome is.\n",
    "\n",
    "**Logistic regression** is the most widely used classifier of this kind. Despite its name, it is a classification method: it predicts the probability that an observation belongs to the positive class. It builds on the linear regression of Chapter 13, so its coefficients can be read in a similar way. This section fits it to the same real shopping session data (UCI Online Shoppers Purchasing Intention, Sakar and others, 2019; CC BY 4.0), training set, and test set as Sections {doc}`1501-knn-scikit-learn` and {doc}`1502-evaluating-classifiers`."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c002",
   "metadata": {},
   "source": [
    "```{contents}\n",
    ":local:\n",
    ":depth: 2\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c004",
   "metadata": {},
   "source": [
    "## The Logistic Curve\n",
    "\n",
    "```{index} logistic function\n",
    "```\n",
    "\n",
    "```{index} sigmoid function\n",
    "```\n",
    "\n",
    "```{index} logistic regression\n",
    "```\n",
    "\n",
    "Linear regression predicts a number by adding up the features, each multiplied by a slope: $z = b_0 + b_1 x_1 + b_2 x_2 + \\cdots$. That number can be anything, such as $-3$ or $250$, but a probability must lie between 0 and 1. Logistic regression fixes this by passing $z$ through the **logistic function**, also called the *sigmoid*:\n",
    "\n",
    "$$p = \\frac{1}{1 + e^{-z}}$$\n",
    "\n",
    "The logistic function turns any number into a value between 0 and 1. When $z = 0$, $p = 0.5$; large positive $z$ gives $p$ close to 1, and large negative $z$ gives $p$ close to 0."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c005",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.056915Z",
     "iopub.status.busy": "2026-11-17T23:43:49.056809Z",
     "iopub.status.idle": "2026-11-17T23:43:49.101530Z",
     "shell.execute_reply": "2026-11-17T23:43:49.101150Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x350 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "def logistic(z):\n",
    "    return 1 / (1 + np.exp(-z))\n",
    "\n",
    "z = np.linspace(-6, 6, 200)\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(6, 3.5))\n",
    "ax.plot(z, logistic(z), lw=3)\n",
    "ax.axhline(0.5, color='gray', lw=1, ls='--')\n",
    "ax.set_xlabel('z (weighted sum of the features)')\n",
    "ax.set_ylabel('probability p')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c006",
   "metadata": {},
   "source": [
    "```{index} predicted probability\n",
    "```\n",
    "\n",
    "Here is the curve in action with a single feature. Fit logistic regression to predict a purchase from the exit rate alone. Then group the training sessions by exit rate and compare the model's **predicted probability** at each group's average exit rate with the actual purchase rate in the group."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c007",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.103032Z",
     "iopub.status.busy": "2026-11-17T23:43:49.102946Z",
     "iopub.status.idle": "2026-11-17T23:43:49.120126Z",
     "shell.execute_reply": "2026-11-17T23:43:49.119697Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>sessions</th>\n",
       "      <th>average_exit_rate</th>\n",
       "      <th>actual_purchase_rate</th>\n",
       "      <th>predicted probability</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>497</td>\n",
       "      <td>0.003</td>\n",
       "      <td>0.296</td>\n",
       "      <td>0.211</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>853</td>\n",
       "      <td>0.008</td>\n",
       "      <td>0.260</td>\n",
       "      <td>0.202</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2041</td>\n",
       "      <td>0.015</td>\n",
       "      <td>0.230</td>\n",
       "      <td>0.188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1535</td>\n",
       "      <td>0.025</td>\n",
       "      <td>0.180</td>\n",
       "      <td>0.172</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1696</td>\n",
       "      <td>0.040</td>\n",
       "      <td>0.096</td>\n",
       "      <td>0.149</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>1225</td>\n",
       "      <td>0.074</td>\n",
       "      <td>0.043</td>\n",
       "      <td>0.105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>222</td>\n",
       "      <td>0.128</td>\n",
       "      <td>0.009</td>\n",
       "      <td>0.059</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>562</td>\n",
       "      <td>0.197</td>\n",
       "      <td>0.005</td>\n",
       "      <td>0.027</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   sessions  average_exit_rate  actual_purchase_rate  predicted probability\n",
       "0       497              0.003                 0.296                  0.211\n",
       "1       853              0.008                 0.260                  0.202\n",
       "2      2041              0.015                 0.230                  0.188\n",
       "3      1535              0.025                 0.180                  0.172\n",
       "4      1696              0.040                 0.096                  0.149\n",
       "5      1225              0.074                 0.043                  0.105\n",
       "6       222              0.128                 0.009                  0.059\n",
       "7       562              0.197                 0.005                  0.027"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "exit_model = LogisticRegression()\n",
    "exit_model.fit(X_train[['exit_rate']], y_train)\n",
    "\n",
    "bands = pd.cut(X_train['exit_rate'], [-0.001, 0.005, 0.01, 0.02, 0.03, 0.05, 0.1, 0.15, 0.3])\n",
    "summary = pd.DataFrame({'exit_rate': X_train['exit_rate'], 'purchase': y_train}).groupby(bands, observed=True).agg(\n",
    "    sessions=('purchase', 'size'),\n",
    "    average_exit_rate=('exit_rate', 'mean'),\n",
    "    actual_purchase_rate=('purchase', 'mean'))\n",
    "\n",
    "by_band = summary.reset_index(drop=True)\n",
    "by_band['predicted probability'] = exit_model.predict_proba(\n",
    "    by_band[['average_exit_rate']].rename(columns={'average_exit_rate': 'exit_rate'}))[:, 1]\n",
    "by_band.round(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "c008",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.121194Z",
     "iopub.status.busy": "2026-11-17T23:43:49.121134Z",
     "iopub.status.idle": "2026-11-17T23:43:49.158655Z",
     "shell.execute_reply": "2026-11-17T23:43:49.158283Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x350 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "grid = pd.DataFrame({'exit_rate': np.linspace(0, 0.2, 100)})\n",
    "\n",
    "fig, ax = plt.subplots(figsize=(6, 3.5))\n",
    "ax.plot(grid['exit_rate'], exit_model.predict_proba(grid)[:, 1], lw=3, label='logistic regression')\n",
    "ax.scatter(by_band['average_exit_rate'], by_band['actual_purchase_rate'], s=60, color='darkorange',\n",
    "           zorder=5, label='actual purchase rate')\n",
    "ax.set_xlabel('exit rate')\n",
    "ax.set_ylabel('probability of a purchase')\n",
    "ax.legend(loc='upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c009",
   "metadata": {},
   "source": [
    "The predicted probability falls along an S-shaped curve as the exit rate rises, from about 0.21 at an exit rate near 0 to about 0.03 at 0.2. (Over this narrow range, only the right-hand part of the S shows.) The actual purchase rates fall in the same direction, but faster than the curve: the curve is too low for sessions with very low exit rates (0.21 predicted against 0.30 actual) and too high for sessions with high exit rates (0.03 predicted against 0.005). One feature alone gives a rough fit; the next subsection uses all three."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "c010",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.159713Z",
     "iopub.status.busy": "2026-11-17T23:43:49.159641Z",
     "iopub.status.idle": "2026-11-17T23:43:49.161234Z",
     "shell.execute_reply": "2026-11-17T23:43:49.160969Z"
    },
    "tags": [
     "thebe-interactive"
    ]
   },
   "outputs": [],
   "source": [
    "### Exercise: The Logistic Function\n",
    "#   1. Write ex_logistic(z) = 1 / (1 + e^(-z)) using np.exp.\n",
    "#   2. Apply it to the array ex_z and print the results rounded to 3 decimals.\n",
    "# Print exactly:\n",
    "# [0.018 0.269 0.5   0.731 0.982]\n",
    "ex_z = np.array([-4, -1, 0, 1, 4])\n",
    "### Your code starts here.\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "### Your code ends here."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c011",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.162030Z",
     "iopub.status.busy": "2026-11-17T23:43:49.161975Z",
     "iopub.status.idle": "2026-11-17T23:43:49.163538Z",
     "shell.execute_reply": "2026-11-17T23:43:49.163297Z"
    },
    "tags": [
     "hide-input"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.018 0.269 0.5   0.731 0.982]\n"
     ]
    }
   ],
   "source": [
    "### Solution\n",
    "ex_z = np.array([-4, -1, 0, 1, 4])\n",
    "\n",
    "def ex_logistic(z):\n",
    "    return 1 / (1 + np.exp(-z))\n",
    "\n",
    "print(ex_logistic(ex_z).round(3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c012",
   "metadata": {},
   "source": [
    "## Fitting LogisticRegression\n",
    "\n",
    "```{index} LogisticRegression\n",
    "```\n",
    "\n",
    "```{index} coefficient; logistic regression\n",
    "```\n",
    "\n",
    "```{index} intercept; logistic regression\n",
    "```\n",
    "\n",
    "Fitting logistic regression with all three features follows the same `fit` pattern as every other scikit-learn model. Behind the scenes, `fit` searches for the intercept and coefficients that make the observed purchase outcomes in the training set as likely as possible."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "c013",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.164841Z",
     "iopub.status.busy": "2026-11-17T23:43:49.164779Z",
     "iopub.status.idle": "2026-11-17T23:43:49.175578Z",
     "shell.execute_reply": "2026-11-17T23:43:49.175272Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Intercept: [-2.441]\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>coefficient</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>page_value</th>\n",
       "      <td>0.088</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>exit_rate</th>\n",
       "      <td>-5.737</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>product_pages</th>\n",
       "      <td>0.006</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               coefficient\n",
       "page_value           0.088\n",
       "exit_rate           -5.737\n",
       "product_pages        0.006"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "logit = LogisticRegression()\n",
    "logit.fit(X_train, y_train)\n",
    "\n",
    "coefficients = pd.DataFrame({'coefficient': logit.coef_[0].round(3)}, index=features)\n",
    "print('Intercept:', logit.intercept_.round(3))\n",
    "coefficients"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c014",
   "metadata": {},
   "source": [
    "**Reading the coefficients.** Each coefficient multiplies its feature inside $z$, so its **sign** tells the direction of the effect, holding the other features fixed:\n",
    "\n",
    "- A **positive** coefficient means that larger values of the feature raise $z$ and therefore raise the predicted probability of a purchase.\n",
    "- A **negative** coefficient means that larger values lower the probability.\n",
    "\n",
    "Here page value and the number of product pages have positive coefficients (sessions that see valuable pages and view more products are more likely to end in a purchase), while the exit rate has a negative one (sessions on pages that visitors often leave are less likely to). These directions match the group averages in Section {doc}`1501-knn-scikit-learn`.\n",
    "\n",
    "The size of a coefficient is harder to read than in linear regression, because the logistic curve is not a straight line: the same change in $z$ moves the probability a lot near 50% and only a little near 0% or 100%. (Statisticians translate coefficients into changes in the *odds* of a purchase; that is beyond this introduction.) The coefficients also depend on each feature's units, so do not compare their sizes across features unless the features have been scaled. The exit rate's coefficient, $-5.737$, looks huge next to $0.006$ for product pages, but the exit rate only ranges from 0 to 0.2 while the page count runs to 705; each extra 0.1 of exit rate lowers $z$ by about 0.57, and each extra 100 product pages raises it by about 0.6.\n",
    "\n",
    "The clearest way to see what the model says is to compare predicted probabilities for sessions that differ in one feature. Here are five sessions with an exit rate of 0.02 and 30 product pages, whose pages have page values from \\$0 to \\$40."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "c015",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.176783Z",
     "iopub.status.busy": "2026-11-17T23:43:49.176701Z",
     "iopub.status.idle": "2026-11-17T23:43:49.180327Z",
     "shell.execute_reply": "2026-11-17T23:43:49.179960Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>page_value</th>\n",
       "      <th>exit_rate</th>\n",
       "      <th>product_pages</th>\n",
       "      <th>probability of purchase</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.02</td>\n",
       "      <td>30</td>\n",
       "      <td>0.086</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10.0</td>\n",
       "      <td>0.02</td>\n",
       "      <td>30</td>\n",
       "      <td>0.184</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>20.0</td>\n",
       "      <td>0.02</td>\n",
       "      <td>30</td>\n",
       "      <td>0.352</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>30.0</td>\n",
       "      <td>0.02</td>\n",
       "      <td>30</td>\n",
       "      <td>0.566</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>40.0</td>\n",
       "      <td>0.02</td>\n",
       "      <td>30</td>\n",
       "      <td>0.758</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   page_value  exit_rate  product_pages  probability of purchase\n",
       "0         0.0       0.02             30                    0.086\n",
       "1        10.0       0.02             30                    0.184\n",
       "2        20.0       0.02             30                    0.352\n",
       "3        30.0       0.02             30                    0.566\n",
       "4        40.0       0.02             30                    0.758"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "what_if = pd.DataFrame({'page_value': [0.0, 10.0, 20.0, 30.0, 40.0],\n",
    "                        'exit_rate': [0.02] * 5,\n",
    "                        'product_pages': [30] * 5})\n",
    "what_if['probability of purchase'] = logit.predict_proba(what_if[features])[:, 1].round(3)\n",
    "what_if"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c016",
   "metadata": {},
   "source": [
    "A session whose pages have a page value of \\$0 has a predicted probability of purchase of only 9%, and \\$40 raises it to 76%. The steps are uneven: going from \\$0 to \\$10 raises the probability by 10 points, from \\$20 to \\$30 by 21 points, and from \\$30 to \\$40 by 19 points. How much a feature changes the probability depends on where the session sits on the curve."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "c017",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.181367Z",
     "iopub.status.busy": "2026-11-17T23:43:49.181301Z",
     "iopub.status.idle": "2026-11-17T23:43:49.182795Z",
     "shell.execute_reply": "2026-11-17T23:43:49.182498Z"
    },
    "tags": [
     "thebe-interactive"
    ]
   },
   "outputs": [],
   "source": [
    "### Exercise: A Two-Feature Model\n",
    "#   1. Fit a LogisticRegression on X_train[['page_value', 'exit_rate']] and y_train.\n",
    "#   2. Print the two coefficients rounded to 3 decimals.\n",
    "#   3. Print whether each one raises or lowers the probability of a purchase.\n",
    "# Print exactly:\n",
    "# page_value: 0.087 (raises purchase probability)\n",
    "# exit_rate: -6.902 (lowers purchase probability)\n",
    "### Your code starts here.\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "### Your code ends here."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "c018",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.183774Z",
     "iopub.status.busy": "2026-11-17T23:43:49.183705Z",
     "iopub.status.idle": "2026-11-17T23:43:49.190795Z",
     "shell.execute_reply": "2026-11-17T23:43:49.190360Z"
    },
    "tags": [
     "hide-input"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "page_value: 0.087 (raises purchase probability)\n",
      "exit_rate: -6.902 (lowers purchase probability)\n"
     ]
    }
   ],
   "source": [
    "### Solution\n",
    "two_features = ['page_value', 'exit_rate']\n",
    "model2 = LogisticRegression()\n",
    "model2.fit(X_train[two_features], y_train)\n",
    "\n",
    "for name, coef in zip(two_features, model2.coef_[0]):\n",
    "    direction = 'raises' if coef > 0 else 'lowers'\n",
    "    print(f'{name}: {round(coef, 3)} ({direction} purchase probability)')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c019h",
   "metadata": {},
   "source": [
    "## Predicted Probabilities\n",
    "\n",
    "```{index} predict_proba\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c019",
   "metadata": {},
   "source": [
    "For each session, the fitted model's `predict_proba` method returns one column per class: the probability of class 0 (no purchase) and of class 1 (purchase). Each row adds up to 1. You usually keep the second column, the probability of the positive class."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c020",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.191938Z",
     "iopub.status.busy": "2026-11-17T23:43:49.191872Z",
     "iopub.status.idle": "2026-11-17T23:43:49.194208Z",
     "shell.execute_reply": "2026-11-17T23:43:49.193898Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.921 0.079]\n",
      " [0.973 0.027]\n",
      " [0.878 0.122]\n",
      " [0.134 0.866]\n",
      " [0.964 0.036]]\n"
     ]
    }
   ],
   "source": [
    "proba = logit.predict_proba(X_test)\n",
    "print(proba[:5].round(3))\n",
    "\n",
    "p_buy = proba[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "c021",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.195225Z",
     "iopub.status.busy": "2026-11-17T23:43:49.195169Z",
     "iopub.status.idle": "2026-11-17T23:43:49.198951Z",
     "shell.execute_reply": "2026-11-17T23:43:49.198680Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>page_value</th>\n",
       "      <th>exit_rate</th>\n",
       "      <th>product_pages</th>\n",
       "      <th>p_buy</th>\n",
       "      <th>predicted</th>\n",
       "      <th>actual</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5218</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.016667</td>\n",
       "      <td>12</td>\n",
       "      <td>0.079</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>421</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.200000</td>\n",
       "      <td>1</td>\n",
       "      <td>0.027</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5110</th>\n",
       "      <td>5.694121</td>\n",
       "      <td>0.042323</td>\n",
       "      <td>34</td>\n",
       "      <td>0.122</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6614</th>\n",
       "      <td>45.711463</td>\n",
       "      <td>0.009091</td>\n",
       "      <td>55</td>\n",
       "      <td>0.866</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1817</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.150000</td>\n",
       "      <td>4</td>\n",
       "      <td>0.036</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5654</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>9</td>\n",
       "      <td>0.046</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4549</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.026500</td>\n",
       "      <td>34</td>\n",
       "      <td>0.085</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12122</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.021839</td>\n",
       "      <td>231</td>\n",
       "      <td>0.245</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       page_value  exit_rate  product_pages  p_buy  predicted  actual\n",
       "5218     0.000000   0.016667             12  0.079          0       0\n",
       "421      0.000000   0.200000              1  0.027          0       0\n",
       "5110     5.694121   0.042323             34  0.122          0       0\n",
       "6614    45.711463   0.009091             55  0.866          1       1\n",
       "1817     0.000000   0.150000              4  0.036          0       0\n",
       "5654     0.000000   0.111111              9  0.046          0       0\n",
       "4549     0.000000   0.026500             34  0.085          0       0\n",
       "12122    0.000000   0.021839            231  0.245          0       0"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_view = X_test.copy()\n",
    "test_view['p_buy'] = p_buy.round(3)\n",
    "test_view['predicted'] = logit.predict(X_test)\n",
    "test_view['actual'] = y_test\n",
    "test_view.head(8)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c022",
   "metadata": {},
   "source": [
    "The `predict` method turns each probability into a label using a cutoff of 0.5: \"purchase\" when the probability is above 0.5, otherwise \"no purchase\" (a probability of exactly 0.5 almost never occurs). You can check that this is exactly what it does."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "c023",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.200197Z",
     "iopub.status.busy": "2026-11-17T23:43:49.200137Z",
     "iopub.status.idle": "2026-11-17T23:43:49.203554Z",
     "shell.execute_reply": "2026-11-17T23:43:49.203192Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "predict matches p_buy > 0.5: True\n",
      "Test accuracy: 0.885\n"
     ]
    }
   ],
   "source": [
    "print('predict matches p_buy > 0.5:', np.array_equal(logit.predict(X_test), (p_buy > 0.5).astype(int)))\n",
    "print('Test accuracy:', round(logit.score(X_test, y_test), 3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c024",
   "metadata": {},
   "source": [
    "A histogram of the predicted probabilities, split by what actually happened, shows how well the model separates the two groups. The vertical axis is on a log scale so that the small bars can be seen. Sessions without a purchase mostly get low probabilities, below 0.2. Buyers are split: some get probabilities near 1, but many get low ones, and only 222 of the 572 buyers (39%) are above the 0.5 line."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "c025",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.204668Z",
     "iopub.status.busy": "2026-11-17T23:43:49.204603Z",
     "iopub.status.idle": "2026-11-17T23:43:49.340808Z",
     "shell.execute_reply": "2026-11-17T23:43:49.340321Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 600x350 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(6, 3.5))\n",
    "bins = np.arange(0, 1.05, 0.05)\n",
    "ax.hist(p_buy[y_test == 0], bins=bins, alpha=0.6, label='no purchase', ec='white')\n",
    "ax.hist(p_buy[y_test == 1], bins=bins, alpha=0.6, label='purchase', ec='white')\n",
    "ax.axvline(0.5, color='k', lw=2, ls='--')\n",
    "ax.set_xlabel('predicted probability of a purchase')\n",
    "ax.set_ylabel('test sessions (log scale)')\n",
    "ax.set_yscale('log')\n",
    "ax.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "c026",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.341847Z",
     "iopub.status.busy": "2026-11-17T23:43:49.341775Z",
     "iopub.status.idle": "2026-11-17T23:43:49.343319Z",
     "shell.execute_reply": "2026-11-17T23:43:49.343063Z"
    },
    "tags": [
     "thebe-interactive"
    ]
   },
   "outputs": [],
   "source": [
    "### Exercise: Likely Buyers\n",
    "#   1. Using p_buy (the test-set purchase probabilities), count the test sessions\n",
    "#      whose predicted probability of a purchase is above 0.7.\n",
    "#   2. Among those sessions, compute the share that actually ended in a purchase\n",
    "#      (use y_test), rounded to 3 decimals.\n",
    "# Print exactly:\n",
    "# Likely buyers: 214\n",
    "# Share who bought: 0.776\n",
    "### Your code starts here.\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "### Your code ends here."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "c027",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.344247Z",
     "iopub.status.busy": "2026-11-17T23:43:49.344181Z",
     "iopub.status.idle": "2026-11-17T23:43:49.346056Z",
     "shell.execute_reply": "2026-11-17T23:43:49.345746Z"
    },
    "tags": [
     "hide-input"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Likely buyers: 214\n",
      "Share who bought: 0.776\n"
     ]
    }
   ],
   "source": [
    "### Solution\n",
    "likely_buyers = p_buy > 0.7\n",
    "\n",
    "print('Likely buyers:', likely_buyers.sum())\n",
    "print('Share who bought:', round(y_test[likely_buyers].mean(), 3))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c028",
   "metadata": {},
   "source": [
    "## Changing the Decision Threshold\n",
    "\n",
    "```{index} decision threshold\n",
    "```\n",
    "\n",
    "```{index} threshold; precision and recall\n",
    "```\n",
    "\n",
    "```{index} cost of errors; threshold\n",
    "```\n",
    "\n",
    "The cutoff of 0.5 is called the **decision threshold**. Nothing forces you to use 0.5. Lowering the threshold flags more sessions: recall goes up (more buyers are caught) and precision usually goes down (more false alarms). Raising it does the opposite. Which threshold is best depends on the costs from Section {doc}`1502-evaluating-classifiers`: a missed buyer costs \\$30 and a false alarm \\$3, both assumptions made for that example."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "c029",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.346986Z",
     "iopub.status.busy": "2026-11-17T23:43:49.346927Z",
     "iopub.status.idle": "2026-11-17T23:43:49.372908Z",
     "shell.execute_reply": "2026-11-17T23:43:49.372502Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>threshold</th>\n",
       "      <th>flagged</th>\n",
       "      <th>FP</th>\n",
       "      <th>FN</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "      <th>error cost ($)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.05</td>\n",
       "      <td>3279</td>\n",
       "      <td>2709</td>\n",
       "      <td>2</td>\n",
       "      <td>0.267</td>\n",
       "      <td>0.174</td>\n",
       "      <td>0.997</td>\n",
       "      <td>8187</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.10</td>\n",
       "      <td>1154</td>\n",
       "      <td>657</td>\n",
       "      <td>75</td>\n",
       "      <td>0.802</td>\n",
       "      <td>0.431</td>\n",
       "      <td>0.869</td>\n",
       "      <td>4221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.15</td>\n",
       "      <td>789</td>\n",
       "      <td>354</td>\n",
       "      <td>137</td>\n",
       "      <td>0.867</td>\n",
       "      <td>0.551</td>\n",
       "      <td>0.760</td>\n",
       "      <td>5172</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.20</td>\n",
       "      <td>627</td>\n",
       "      <td>240</td>\n",
       "      <td>185</td>\n",
       "      <td>0.885</td>\n",
       "      <td>0.617</td>\n",
       "      <td>0.677</td>\n",
       "      <td>6270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.30</td>\n",
       "      <td>476</td>\n",
       "      <td>160</td>\n",
       "      <td>256</td>\n",
       "      <td>0.888</td>\n",
       "      <td>0.664</td>\n",
       "      <td>0.552</td>\n",
       "      <td>8160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.40</td>\n",
       "      <td>370</td>\n",
       "      <td>102</td>\n",
       "      <td>304</td>\n",
       "      <td>0.890</td>\n",
       "      <td>0.724</td>\n",
       "      <td>0.469</td>\n",
       "      <td>9426</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.50</td>\n",
       "      <td>298</td>\n",
       "      <td>76</td>\n",
       "      <td>350</td>\n",
       "      <td>0.885</td>\n",
       "      <td>0.745</td>\n",
       "      <td>0.388</td>\n",
       "      <td>10728</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   threshold  flagged    FP   FN  accuracy  precision  recall  error cost ($)\n",
       "0       0.05     3279  2709    2     0.267      0.174   0.997            8187\n",
       "1       0.10     1154   657   75     0.802      0.431   0.869            4221\n",
       "2       0.15      789   354  137     0.867      0.551   0.760            5172\n",
       "3       0.20      627   240  185     0.885      0.617   0.677            6270\n",
       "4       0.30      476   160  256     0.888      0.664   0.552            8160\n",
       "5       0.40      370   102  304     0.890      0.724   0.469            9426\n",
       "6       0.50      298    76  350     0.885      0.745   0.388           10728"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rows = []\n",
    "for threshold in [0.05, 0.1, 0.15, 0.2, 0.3, 0.4, 0.5]:\n",
    "    flagged = (p_buy >= threshold).astype(int)\n",
    "    tn, fp, fn, tp = confusion_matrix(y_test, flagged).ravel()\n",
    "    rows.append({'threshold': threshold,\n",
    "                 'flagged': tp + fp,\n",
    "                 'FP': fp, 'FN': fn,\n",
    "                 'accuracy': round(accuracy_score(y_test, flagged), 3),\n",
    "                 'precision': round(precision_score(y_test, flagged), 3),\n",
    "                 'recall': round(recall_score(y_test, flagged), 3),\n",
    "                 'error cost ($)': error_cost(y_test, flagged)})\n",
    "\n",
    "thresholds = pd.DataFrame(rows)\n",
    "thresholds"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "c030",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.373969Z",
     "iopub.status.busy": "2026-11-17T23:43:49.373899Z",
     "iopub.status.idle": "2026-11-17T23:43:49.399724Z",
     "shell.execute_reply": "2026-11-17T23:43:49.399385Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAnwAAAFkCAYAAAC6gOY9AAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjExLjIsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvgI3uAAAAAAlwSFlzAAAPYQAAD2EBqD+naQAAY+xJREFUeJzt3Qd4k2XXB/B/955Ayywgew9l7y0bVHCgOHGACCgqjvfz1VfB8SoqbhQExe0rewiyNyJ7701bSvdu0u86d3nSJE1L26Rtxv93Xb3SPnmaPM3TJCf3fZ9z3BISEnJBRERERE7LvaIPgIiIiIjKFgM+IiIiIifHgI+IiIjIyTHgIyIiInJyDPiIiIiInBwDPiIiIiInx4CPiIiIyMkx4CMiIiJycgz4iIiIiJwcAz4iIiIiJ8eAj5xGRkYGTp8+rS7JfvE82T+eI8fA82T/MuzofYkBHzkVnU5X0YdAxcDzZP94jhwDz5P909nJ+xIDPiIiIiInx4CPiIiIyMkx4CMiIiJycgz4iIiIiJwcAz4iIiIiG9Ppc7ElOhurYj3UpfxckTwr9N6JiIiInMzis+mYtiMBl9P0AHyAY0mo7p+CtzuEYlgdvwo5Jo7wEREREdkw2Htw3fUbwV6+K2l6tV2urwgM+IiIiIhsQKZtZWTP0uSttu2lnYkVMr3LgI+IiIjIBrZGZxUY2TMmYd6lVJ3ar7wx4CMiIiKygdNJOcXaLzq9/LtvMGmDiIiIyArXMnT47FAKPj+UUqz9I/08UN4Y8BERERGVQnSaDrMOpmDOsVSk5dx8XZ4bgOoBHugc6Y3yxoCPiIiIqARkHd5HB5Ix/3gqMoxmZ73cga5VvbH+ct4avVyzYE/MaB8CD3ftp/LDgI+IiIioGM6n5ODD/Sn4/kQqsoxyM3w8gLENAzCpeSBqBnqa1eHLIyN7EuxVVB0+BnxERERERTiTlIP39yfjp5NpMJ659fd0w8ONAjCxeSCq+uevy5OgbnCUL9ZfSMGRS7FoUqMKetYKrJCRPQ0DPiIiIiILjidkq0Dvt9Pp0BkFeoGebhjXJAATmgeisq/lBAwJ7rpEeiEqS4dakV4VGuwJBnxERERERg7HZ+O/+5Lxx5l0k3V4wd5ueKJJIMY3C0SYj2NVtmPAR0RERARgX1yWCvSWnMsweTzCfNwwvmkgxjUJRKiDBXoaBnxERETk0nbHZuHdfclYdcE00Kvs667W5z3SOABBkoLrwBjwERERkUvaFp2J9/YmY+3lTJPtVf3cMbFFEB5q6I8ABw/0NAz4iIiIyGXk5uZi09UsvLc3SV0aq+HvgcktA/FAgwD4elZskoWtMeAjIiIilwj01l7OG9HbHmMa6EUFeuDZlkG4t74/fDycK9DTMOAjIiIipw70Vl3MUIHe7mvZJtfVC84L9EbX84dXBZdNKWsM+IiIiMjp6HNzsfRchsq63X/dNNBrFOKJqa2CMLKuHzydPNDTMOAjIiIip6HT52LR2XQV6B1OyDG5rlmYJ55vFYxhdXzh7uYagZ6GAR8RERE5vBx9ruqIIZ0xTiSaBnqtK3nh+VZBGBjleoGehgEfEREROawsXS5+OpWGmfuTcSZZZ3JduypeeKF1MPrW8IGbiwZ6GgZ8RERE5HAydbn4/kQqZu5PwcVU00Cvc6Q3XmgdhB7VGOhpGPARERGRw0jPycW846n4+EAyLqfpTa7rWd1HTd12qepTYcdnrxjwERERkd1LydZj7tFUzDqUgph000CvXw0fPN86CO0jGOgVhgEfERER2a2kLD2+PpqKTw6m4HqmaaA3KMpXjei1qexdYcfnKBwm4MvKysKWLVuwdu1apKSkYMSIEejRo0eRv7N+/Xps3rwZaWlpaNSoEYYPH47Q0FCH2Z+IiMhVJWTq8cXhFPWVkJVr2C6pF8Pr+OG5VkFoEe5VocfoSBwi4Hv99dfx9ddfIzk52bCtQYMGhQZ82dnZePDBB7F8+XKT7TNmzMAvv/yCli1b2vX+REREriouQ4fPDqVg9pFUJGXnB3pSH/nOunmBXuNQBnpOGfDt3LkTGRkZ6NmzJ7y9vfHnn38Wuf/06dNVcBUWFoaHH34Y4eHhWLJkCXbs2IH7778fu3btgo+Pj93uT0RE5Gpi0nVq2vabo6lIzckP9KS17d31/PFsy0DUD2GgV1puCQkJ+Y+qnZKASKZAg4OD8dVXX+GFF15QQdT48eML7CvTpfXr11ejahs2bEDTpk3Vdp1Op6ZQZUr1008/xZgxY+xyfyo9+VBw4cIF1KpVC76+vnwo7RTPk/3jOXIMznKerqTp8NGBZMw7loZ0XX5I4uUO3FffH1NaBqFOkEOMT9n1OXKHA2jXrp0K9opj06ZNKsgaOHCgIbgSHh4emDJlivp+5cqVdrs/ERGRK7iQkoOp2xLQ6ter+OJwqiHY8/EAxjUOwD93RuKjLmEOG+zZG6sfxUOHDqlRKkmoOHXqFOLi4pCYmIiQkBBUqlRJjW516dIFXbt2NQl4ysrhw4fVZceOHQtcp23T9rHH/YmIiJzZ2eQcfLA/GT+eTEO2UdKtn4cbHm7sj4nNg1DN36MiD9EplSrgS0pKws8//4xvv/1WBXyWpKen4+rVq+r6RYsWqW0tWrRQyQujRo1SAWFZiI2NVZc1atQocF1AQIDKitX2scf9ixoWpptnchtfkn3iebJ/PEeOwdHO08kkHT46lI7/ncuE0cwt/D2Bhxv44snGfqjiKxOP2cjIyIYzyCrDc1TSKeISBXxywLKG7v3330d8fLzaVrNmTbRt2xa33nqr+l4SE4KCglRGrewjc9e7d+/GP//8gwMHDmDq1Kl466238Nxzz+Hxxx9XSRi2pD2oXl6WF3ZKckRCQoLd7l+Yy5cvq3V/dHPR0dF8mBwAz5P94zlyDPZ+nk6lumHuRS+sjvWAXhVVyRPgkYu7q+fg3urZCPVKQ0YscAHOKdrG50iWhd1yyy1lF/DddtttOH/+PCIiIlTCxOjRo9G6deti//6ePXtUGZLff/8dr776KmbPno19+/bBlvz8/AwjjJbIdm0fe9y/MNWrV7/pPq5Ogmt5UkVGRtr8gwTZDs+T/eM5cgz2fp4Oxudg5qF0LLtgOroV6u2GcY188VhDX4R4O0QqgVOcoxIFfDk5OXjnnXfUtGxpsk3atGmjvl577TXMmzcPH330EWytWrVq6vLcuXMFrpMRR5mONl5LaG/7F6ais3sciTyp+HjZP54n+8dz5Bjs7TztuZaFd/cmY8UF06VIlXzc8XTzQDzaOADBTh7o2eM5KtEjLiN0TzzxhNUHLb8vtyO3Z2taEWMpgWJOunRoawntdX8iIiJHtCM6E3f9eQ29lsSaBHuRfu54s10w9o+KVCVWXC3YsxcletRtXRy4LIoNS+ZrlSpVVIC1YsUKw3ZZJyejk0Lq39nr/kRERI5k89VMDFt5DQOWX8OaS5mG7TX8PfBuhxDsvasqnm4ehAAprEcVxiEKL69evdrQluzo0aPYtm0bOnToYJgKHTRoEPr162fYf86cOXj22WfVosY+ffqoRBLpYyvz6FLTTzp1uLm52e3+5PgFLqlwPE/2j+fIMVTkecrNzcX6y5l4d18ytkWbrtGrFeiBZ1sE4b4G/vCRNhkuLMOO3pccoprh3r17MXfuXJNt0pZMvrSEBuOA75FHHlElYSSb2LgNW/v27fHdd98VCK7sbX8iIiJ7JIHenxcz8d6+JPwda1o65ZYgDzzbKki1QfOSxrfk/CN8UpLlm2++UWv0/P39cfvtt1s1bSkBn5R1KYyUhbGULSxB1vbt21Wni4YNG6os46LY2/7kuJ+kqHA8T/aP58gxlOd50ufmYvn5DLy3Lxn74kwDvYYhnpjaKgh31PWDJwM9u30ulSrgk0LBkyZNUuVZPvzwQ5PrUlJS0L9//wLdI6RXrPSMJXKFJxYVjufJ/vEcOYbyOE86fS4Wn0tXgd7h+ByT65qGeeL5VkEYVtsPHgz07P65VKoVlNJlQ9bUWTp4CQAl2JMizP/5z38wbtw4uLu7Y8GCBewZS0RE5ABy9Ln45VQaOi2MwcPr402CvVaVvPB973BsHh6BkXX9Gew5iFKt4dMSKAYOHFhgbv+HH35Qa9IkKGzWrJnaLtO6UnNPgj6Z3iUiIiL7k63Pxc+n0vDBvmScTjbt7nRbFS883yoY/Wv6cO25qwR8x44dU5f169c32X7q1CnVAkxqz2nBnrjvvvtUwFcWdfeIiIjIOpm6XPxwIg0zDyTjfIppoNcp0hsvtApCz+oM9Fwi4Bs8eLDh+7i4OHUp07XGGaZaT9grV66Y7C8jf0KCQePty5Yts/b4iYiIqJTSc3Lx3fFUfHQgBZfSTAO97tV88ELrIHStavuauWTHAd/YsWMN3+/atUv1h7vnnntMiif/+OOP6nLEiBEFMlB37typgkNJ3pBLliIhIiKqGKnZesw9lopZB1MQna43ua5vDR+VjNEhkoGeSwZ8d999t+H7mTNnqgLIUVFR6Nmzp2H7rFmz1OWTTz6JevXqmfSMlT68sr9M7xIREVH5S87W4+sjqfj0UAquZZgGegNr+apAr20Vb54aJ1SqNXxS5FgCvpdfflkVDpZA7osvvsDBgwdV9wvjYE9ISrKQtGQiIiIqXwmZenx1JAWfH05BfKZpNbZhtX1VHb2WlRjoObNSBXwTJkxQGbdSfuXWW29VZVf0+rxPChIEmpM2YqJ79+7WHi8REREV0/UMHT47nIqvDqcgKTs/0JOyeVIo+bmWQWgS5sXH0wWUKuCrWrUqli5diqlTp6q+thLsRUZG4pVXXsGQIUMK7P+///3P0POWiIiIylZsuk5N28r0bUpOfqAnrW1H1/PHcy0DUT+EgZ4rKXUvXZm6lXp8qampSE9PR6VKlQpNxJDafJKpW6NGDWuOlYiIiG50wNgSnY0jsR5o4p2NnrV8VAHkq2k6fHwwGXOPpiFdlx/oebkD99b3x7Mtg1AnqNRv/eTArD7rAQEB6qso1atXt/ZuiIiICMDis+mYtiMBl9NkKZUPcCwJkX4paBHuhU1XM5FpVF3F2x0Y2zAAk1oEolYgAz1XxrNPRETkQMHeg+uuwzTtAqq0SvSlTMPPvh7AQ40C8EzzIFQP8Cj34yQHD/h0Oh08PGz3j2Pr2yMiInLmaVwZ2TMP9ozJwqoJzQLwTIsgRPjx/ZXyuaMEOnbsqHrkSqBmDfl9KdLcoUMHq26HiIjIVWyNzroxjVs4CQYH1PJjsEfWBXzSOu2JJ55QXTTeffddnDlzpiS/jhMnTmD69Olo06YNnnrqKSQmJpbo94mIiFzR1quZeGlHXvvSm4lOt25QhpxTiaZ0//nnH3z00Uf47LPPVOAmX5KtK7X4JIirWbMmwsLCVBJHSkqK6rAhRZd3796tvo4fP65uR65//vnnMWnSpLL6u4iIiByaVLdYfTETMw8kY1t0VrF/L5JTuWSBW0JCQlHLASyKiYnBvHnzVJeN8+fPF/v36tSpgwceeED15a1SpUpJ75aoSBkZGeoDhnR08fX15aNlp3ie7B/PUcWv1Vt8Lh3v70/BwevZJtdJHT2jaisF1u9Jgsb+uyJViRaqeBl29L5UqizdiIgINUL33HPPYfv27di8ebMqwHzq1CnExcWp2nyBgYEIDw9H/fr10alTJ3Tt2lWtASysVh8REZEry9Ll4qdTafjoQDJOJZlOyzYO9cSUlkGqzMoj6+PVNuO4T3tnndE+hMEe2b4si7RU69y5s/oyxuxbIiKi4knN1mP+8TR8cjAFl9JMA722lb1UseRBUb5wvzFg4uHmZlSHL4+M7EmwN6yOHx92Kr86fCy1QkREVLSETD1mH0nB54dTcT3TNPu2ezUf1f5MLs1nxiSoGxzli/UXUnDkUiya1KiCnrUCObJHRWLhZSIionIUnabDZ4dSMOdYKpKzTRfkyUiejOjdVsW7yNuQNXpdIr0QlaVDrUgvBnt0Uwz4iIiIysG55BzMOpiC706kmrQ/k/yKu+r6YXLLIDQN8+K5oDLBgI+IiKgMHU3Ixsz9yfjtdLpJhq0kYNzfQLpiBKJOEN+OqWzxP4yIiKgM/BObhQ/2J2Pp+QyT7YGebni4cQDGNwtENX+2P6PywYCPiIjIhsWSN13NC/TWX840uS7Mxw1PNg3E400CEeZTokZXRFZjwEdERGQlfW4uVl7IUFO3u2JNiyVX83fHhGaBeKhRAAK9GOhRxWDAR0REVEo5+lz8cSZdBXqHE3JMrqsb5IHJLYJwT31/+EiLDCJHDfhmzZqlLidOnGjTfYmIiOxZRk4ufjyZho8OJuNssmmx5KZhnqq0yog6fvBkizNy5F66mtDQUHWZkJBg032JHL1nIRWO58n+8RwVLiVbj7nHUvHpwRRcTTctlty+ijeebRWIATV9y6WNKM+T/cuwo/clz/JaxEpEROSormfo8OWRVHx5OAUJWabvab2r++DZVkHoEunNfvFkt8ol4IuJiVGXgYGB5XF3RERENnElTadG82RULzUnP9CT8bshtfO6YrSpXHRXDCKHC/j27NlTou0iKSkJ8+fPV983bNiwpMdHRERU7s4k5eCjA8n44WQasoxmbiX3YnQ9f0xuEYhGoeyKQU4a8PXq1atE2809/PDDJbk7IiKicnXoejZmHkjG/86kQ280c+vrATzQMAATmwciKpAFLsjxlOi/VhYdGpOFiJa2a2TRqr+/P+rVq4fRo0dj+PDh1hwrERFRmdgVk4X39yerWnrGgrzc8FjjADzVLBARfuyKQS4S8B04cMBi5q35diIiInsnCYXSDUMCvc1Xs0yuq+TjrlqfPdo4AKHsikFOwKpx6Y8//th2R0JERFROXTGWnc9Q7c/2XDPtilEzwANPNw/E2Ib+8PdkVwxyHlYFfGPHjrXdkRAREZWhbH0ufjudjg/3J+NYomlXjPrBnpjcMhCjb/GHN7tikBPytHY4/MiRI3B3d0fjxo1Nrtu2bZvqrnHx4kU0b94cr7zyCmrUqGHt8RIREZVIek4uvj+Rio8PpuBCimlXjJbhXqq0ytDavvBgVwxyYlYFfOvXr8fIkSPRp08f/P7774bt27dvx5AhQ6DT5T2x9u/fj02bNqkvbd0fERFRWUrK0mPO0VR8eigFsRmmXTE6RXrjuZZB6FPDh8WSySVYtUDhu+++szi1+9///lcFe48++ii+/PJLREVFqYzezz77zLqjJSIiuolrGTq8uTsJzX+9in/vTjIJ9vrX9MGKQZWxYlAV9C2nFmhEDj/Cp2XntmzZ0rAtJSVFjfw1aNAA77//vtoWHByMe++9FytXrsTLL79s7TETEREVcDElB58cSsG8Y2lI15l2xRhRx0+t0WtViV0xyDVZFfBdunRJXVavXt2w7e+//0ZOTg769etn2NaxY0d1efr0aWvujoiIqICTidn48EAKfj6VhmyjmVsvd+Ceev6Y1CIQ9UPYFYNcm1UBn16f98xKTU2Fj4+PIVlDtGvXzrCfn5+fuszONk1/JyIiKq19cVmYuT8Fi86mw6gpBvw83PBgI3883SwQNdkVg8j6gK9OnTo4evSoStIYNGiQ2rZ8+XKTUT1x5cqVAiOBREREpbH1aiZm7k/G6kuZJtuDvd3weJNAPNk0AJWlFxoR2SbgGzBggAr4nn/+ecTGxuLgwYNqXV+HDh1QrVo1w35nzpxRl/Xr17fm7oiIyEVJGbA1lzJVseRt0aZdMSL83DGhWSAebhSAYG8WSyayecA3adIkLFq0CGfPnlXfC5naff311032W7FihbocOHCgNXdHREQuRqfPxeJz6fhgfwoOXDddFlQr0AOTmgdiTIMA+Hky25aozAK+8PBwlZE7Z84cNdJXtWpV3H///WjYsKHJfjExMWqKd/DgwdbcHRERuYgsXa5KwvjwQDJOJZkWS24U4okpLYNw5y1+8GKxZKKyD/iEFFJ+9tlni9xn3rx51t4NERG5gNRsPeYfT8MnB1NwKc000GtTOa8rxuAoX7izfh5R+QZ8RERE1krI1OPro6n4/FAK4jJNu2J0q+qN51oFoUc1dsUgqtCATxbTrl27FqtWrcKJEydU8eXAwEBVfFkSO3r37s1q5kREVEBMug6fHUrBN0dTkZxtXFwFGFjLV43otYtgsWSiCg/4zp07p1qoScFlc+vWrcNXX32F9u3b45tvvkGtWrWsvTsiInIC55Jz1LTtdydSkWE0cytL8u6s64fJLYLQLJzFkonsIuBLTEzE8OHDVZau6NWrF9q2bYvIyEhcvXoVe/bsUUHfzp07MWLECPW9tFkjIiLXdCwhW9XQ+/V0Ooy6n0GqqYxp4I9nmgehbjBXGxHZmlXPqk8++UQFexEREViwYIFJdw3Njh07VObuqVOn8Omnn+Kll16y5i6JiMgB/RObpWroLTufYdIVI8DTTdXPm9A8ENX8WSyZqKxYVaFy6dKl6vLdd9+1GOwJKcL8zjvvqO+XLFlizd0REZEDkfXdm65kYuSqa+i9NBZLjYK9UG83TGsdhAOjIvFm+xAGe0T2PMKnTeX27du3yP369etnsj8RETl3oLfyQoYa0dsVa1osuap0xWgeiIcaBSDIi10xiBwi4HO7UQdJrzdNoTenXa/tT0REzidHn4s/zqRj5oFkHI7PMbmuTpCHSsS4t74/fDz4XkBU3qz6eFW3bl11uXLlyiL3k3ItxvsTEZHzyNTl4ttjqbjtf9EYtzHeJNhrGuqJ2d3D8PcdkWpUj8EekQOO8A0bNgyHDh3CtGnTVGZujx49CuyzadMmQ6KGZPQSEZFzSMnWY+6xVFVH70qa6UxPuyp5XTEG1GJXDCKHD/jGjx+PX3/9FSdPnlTBnCRotGnTRgV/0dHR2LdvH7Zt26b2lf66Tz31lK2Om2zQkHxrdBai03WI9PNA50hveLAnJREVQ3ymHl8eTsGXR1IQn2laLLlXdR8V6HWt6s1lPETOEvAFBQVh0aJFGDduHLZu3apKsMiXua5du2L27Nmq+wZVvMVn0zFtRwIuG30ir+7vjrc7hGJYHb8KPTYiqvgPg1uis3Ek1gNNvLPRs5aP4cPglTQdPj2Yokb1UnNMA70hUXldMdpWYVcMIntkdXXLGjVqYPny5diyZQtWr15doLVa//790blzZ9scLdkk2Htw3XWTOlhCpmNk+7xe4Qz6iFyU6YdBH+BYEqr7p6g+tgevZ2PBiTRkGc3cSu7FqFv8MLllEBqHsisGkT2zWTnzLl26qC+y70/u8mJuHuwJ2Saf4V/amYjBUb6c3iVyMYV9GJTg77ltiSbbfDyABxoEYGLzQNQOYlcMIkfAZ6oLkTV7xtO45uSF/lKqTu3XrZpPuR4bEdnnh0FjgZ7AY00C8VTTQESyKwaR6wR8Ul9PEjPc3d3RqlWrQveTfWRf2Uf2LS/nz583TDHXrl0bzZo1g5dX4dMO8fHxag1ienq6SjKR/YtS1vvbmiRo2HI/InKND4Oa2T3CMTCK63yJXC7gk/p69957LwYMGICff/650P2mT5+u9v3ll1/Umr6yFhsbi+eff14llEjFd80tt9yi2sBZ6gwyc+ZMzJgxA1lZWYZt3bp1w9y5c1G5cuVy378sSDauLfcjIudQ3A955okaROQ4rBpumzdvnrp86KGHitzvwQcfNNm/rMn9LVy4ED4+Piq4k5IxMsJ3+vRpFaAePXrUZP/58+fj9ddfR05OjgrCRowYgfDwcFVD8IEHHihw+2W9f1mR0iuSjVtYjXvZXiMgr0QLEbmOEwmmXTEKww+DRC4a8B08eFBdtmjRosj9tOulSHNZk369UiImNDQU27dvx2+//aYCzX/++Qd33HEHsrOzVe1AjU6nUyOQQkbblixZgm+//Ra7du1CvXr1VB1BrVNIeexflqS0gpReEYUFfTPahzBhg8hFJGfr8dSmeLyzL7nI/fhhkMjFA76YmBh1KYWWi6Jdf+XKFZS1xMS8bLJOnTqhTp06hu0eHh645557TPYRsqbu6tWrKsPYuBNIpUqV8OKLL6rvZWq4vPYva1JnT0qvVPM3PfXe7mBJFiIXsismC90WxeDHk2km280/DGo/88MgkQsHfMHBwery0qVLRe538eJFdVkehZdl1Ezu58iRI8jIyDC5bvfu3eqyZcuWJgklonfv3gVuq0+fPib7lMf+5RX0HRhVFUtur4yaAXn/AlJbq3Vl1tEicnY5+ly8szcJty+PxdnkvLV7QV5u+KJbGOb3CivwYbB6gAc/DBK5etJG69atsWbNGpWM8cILLxS6n1wvisrktRUJ9t58801MmTJFrd8bNWoUAgIC1BSqTO/KyJ820mc86mg8Gmg8CifdRC5fvlxu+xfGPHi1hXZhwP31fPD2/nT18+8nkvFUE8fNwNMSYowTY8j+8DxVnHMpOjy9LQW7ruWv2WtX2ROfdApE7cC8QK/P0FBsvpKOE9EJaBAZiq7V/NQyj7J4DSLr8Lnk2ufI19e3/AI+SYCQgO+///2vKjEyePDgAvssXrwY77//vmH/8iBJJBLkTZo0Ca+99pphu0ypfvbZZ/D2zk9KSEvLm87w87Mc6MjtSDmV8tq/MBIUynpAW7vNSyZs8o7tt5PJGBJ4DY5O+jiT/eN5Kj9SrGBFrAfePeWNVF3eJK0HcvFYVDYeqpUG9/gkXDB6GbpFvqpI7a04XC56AofsAJ9LrneOPDw8VOWRcgv4Ro4ciR9++AF//fUXxowZgw4dOqi+uZIwkZCQoLJQd+7cqfbt168f7rzzTpQHGeGTIDQkJESVjPH398exY8fUWrkzZ86opA1tXaEW/EkyhyWZmZkq21dT1vsXpnr16igLtQA0P5WAgwk6HErxQG5YdUQFOmZZFvkEJU8qObfGQT3ZF56n8pWYpceLu1Kx8Hz+CIOM5n3WKRC3FrKMg+fIMfA82b8sO3pfsirgkyLKUnJk8uTJKoiSBAX5Mnf33XerOnTlUXR58+bNKtjr2LGjmsI1Xjc4Z84cPPvss3j11Vcxe/ZstU2rgWdpWlVG5yRwrVu3rmFbWe9vq6HbkrijXgAO7k5S36+8osczLQLgyORJVZaPF9kGz1PZ23w1E09uTMLF1PzZgfvq++OdjiEI8rr56zHPkWPgebJ/3nbwvmR1BCZTkhI8bdy4URU7Hjp0KHr27KkuZV3fli1b8OWXX6pRtvIgAZ+47777CiSJyFSvPOgy8qhp0qSJutRGIo3JNincrO1THvtXhBF18qeb/zibt56PiBxXli4Xb+xOxNAV1wzBXoi3G77tGY7PuoUVK9gjIudis166kvlqnP1aUbTpUalvN3bs2AIBlgyvGg+rSiFkWV+3bNkyHD9+XLU8ExKIffjhh+p7mRYur/0rwi3BnmhVyQv74rKx51o2zibnoA4bohM5pBOJ2Ri3IR574/KXkXSt6q2ycGtKM1wicklO9+yX1m1vvPEGfvzxR5UhK6VPZHRRumv89NNPap+BAwca9pdRwMceewyzZs1SSSfjxo1TXTBkvZ+MBNasWVNl+pbX/hVlZB0/FfCJhWfSMbllUEUfEhGVgHyInH88DS/tTETajRZoMpD3attgPN0skAXViVycW0JCgtM1R5QpZFmnZylRQkbcFixYYKghqCVOSNKJZBybl02RtYlt27Y12V7W+1cEGdVr/VteFpGM9m0YFgFHI2UjLly4gFq1alX4WgkqHM+T7cVl6PDMlgQsO59fOqVBiCdmdw9D68olXyjOc+QYeJ7sX4YdvS85ZcAnzp8/j6VLl6r+uTKNGxERoTKIe/ToATc3N4ufjleuXKnWAKanp6NBgwYYPXq0CsosKev9K0KvJTFqSlfsuTMSdYMdawDYnp5YVDieJ9taeykD4zfF42q63rDtkUYBeLN9MPw9S7dWj+fIMfA82b8MO3pfctqAj0ru4wPJ+L+/87J1X7s1GFMcbFrXnp5YVDieJ9vIyMnFG/8k4rNDqYZtlXzcMatrKAZFWVdAnefIMfA82b8MO3pfYqoWmbRc0/xxhtm6RPbqSHw2+iyNMQn2+tTwwdYREVYHe0TknBjwkYFk5ra9UYh1//VsnErMb79ERBVPloZ8eTgFPZfE4FB83vPTxwN4u0MIfu1XCZH+jlk0nYjsPOD76quv8MUXX9juaMgusnU1C1mTj8huRKfpMGp1HF7ckYjMG3WUm4Z5Yu2QCDzZNBDuFtYmExHZJOB78cUXMW3aNGtuguzM8Loswkxkb1ZeSEeXRTFYcynTsO2ppgEq2GsWbrk9GhGRzQK+sLAwdSlZsOQcogI9cVuVvDeQg9ezVRFXIqoYaTl6PLctAfesuY5rGXlZuJF+7vi9fyXM6BAKX0+O6hFROQR8rVu3VpeHDx+25mbIzhi3WpMizERU/vbFZaHn4lh8czQ/MWNQlC+2jIhAnxrMQieicgz4xo8fry7feust5ORwgb+zGM7eukQVRp+bq0ok9V0ai+M3Eqf8PNzwYedQLOgdjsq+TMwgopKzqrJuVFSUahU2e/ZsdO7cWQWA0iu2cuXKhf6O1kuW7FetQE+0r+KNnbFZOByfg2MJ2WgUynVCRGXtUqoOT22Kx8Yr+Wv1pPONdMxoyOcgEVVUwNe+fXvD98ePH8fkyZNv+jsJCQnW3CWVkxF1/VTAp2XrvtiaAR9RWVp0Nh2TtsQjISuvFr6szpvcIhAvtQmGtwfX6hFRBQZ84eHhVt492fO07ss7Ew3r+F5snd97mIhsJzlbj2k7ErHgRJphWw1/D3zRPQzdqvnwoSaiig/4pE8tOacaAR7oGOGN7TFZOJKQoyr7NwnjKB+RLf0dm4VxG67jTLLOpBbmzM6hCPVhXXwish2+olCR07oaFmEmsp0cfS7e3ZuEActiDcFeoKcbPu8Whjk9wxjsEZHNMeCjQg2r7afWEQmWZyGyjbPJORiy4hqm70mGLm+5HtpV8cKm4RG4t74/3Ngxg4jsbUrXuL/j2rVrsWrVKpw4cQIpKSkIDAxEgwYNMGDAAPTu3ZsvYg6oukzrRnpjW3QWjiVyWpfIWj+fSsPUbQlIzs6L9NzdgBdaBWFqqyB4yg9ERPYa8J07dw6PPvoo/v777wLXrVu3TvXblWzeb775BrVq1bL27qgCijBLwCf+OJvOdXxEpZCQqcfU7Qn47XR+IfPagR74qnsYOkQyMYOI7DzgS0xMxPDhw3H27Fn1c69evdC2bVtERkbi6tWr2LNnjwr6du7ciREjRqjvg4OZ7elIhtXxUxmEuTemdV9qHcTRWqIS2HI1E09sjMfF1PzEjHvq+eHdjqEI9uaqGiJygIDvk08+UcFeREQEFixYgHbt2hXYZ8eOHbj//vtx6tQpfPrpp3jppZesuUsqZ9X8PdAp0htbo7NU1X8pxMxm7UQ3l63Pxdt7kvDB/hT1gUkEe7vhw06huOMWfz6ERFSurPp4uXTpUnX57rvvWgz2RIcOHfDOO++o75csWWLN3VEFGWmUrSvTukRUtFOJOSoD932jYK9LVW9sGR7BYI+IHC/g06Zy+/btW+R+/fr1M9mfHDtbV5J0iKggeW7MP56Kbotj8M+1bLXN0w147dZgLB5QWbUtJCKqCFa9+mjlA/R6fZH7adez3IBjivT3UKMTm69m4WRSDg7G56BFOIswExm7nqHDM1sSsPR8hmFb/WBPfN0jDK0re/PBIiLHHeGrW7euuly5cmWR+0m5FuP9ybGndReeyW8BRUTA+ssZ6LIoxiTYe6ihPzYMq8Jgj4gcP+AbNmyYupw2bRo2bNhgcZ9NmzYZEjUko5cc09DafqpmmPiD07pESqYuF6/sTMSIVXG4kpY3kxHu444FvcPxYZcwBHgxC5eInGBKd/z48fj1119x8uRJFcxJgkabNm1UWZbo6Gjs27cP27ZtU/s2bNgQTz31lK2Om8pZhJ8Hulb1wcYrmTidrMP+69loVYnTVOS6pL/0uI3xOHg9b62e6F3dB591C0NVf48KPTYiIpsGfEFBQVi0aBHGjRuHrVu3qhIs8mWua9eumD17tuq+QY5LmrpLwKclbzDgI1dNzJh9JBX/93ciMm6U1pNyev++LQRPNg2AO1ujEZEdsjplrEaNGli+fDm2bNmC1atXF2it1r9/f3Tu3Nk2R0sVamgdXzy3HdDn5pVn+b9bg5mIQy4lJl2HpzfH48+LeR98RJNQT8zuEY7mTGQiImcN+KKiotSn3QsXLqBLly7qi5xXZV8PdK/mg/WXM3E2WYd9cdlckE4uY9WFDEzYHI9rGflVCZ5oEqBG9vyk9goRkR2zakVxdnY2kpOTbXc05BDTupqFLMJMLiAtR4+p2xJw95o4Q7AX4eeO3/pVwjsdQxnsEZHzB3w1a9ZUl9euXbPV8ZCdG1LbFx7M1iUXsT8uC70Wx+Lro6mGbbfX8sXWERHoW9O3Qo+NiKjcAr5Bgwapy8WLF1tzM+RAKvl6oEc1H/X9uRQd9sblZygSOQt9bi5mHUhGn6WxOJaYo7b5ebjhg06h+LFPuFreQETkMgHfc889h0aNGuFf//oXFixYcNOOG+QcRhj31j3D3rrkXC6n6jByVRz+9XcSsm+8pLUM91JFlB9pHMBEJSJyvaSNp59+GlWrVsWxY8cwYcIEvPHGG6hXrx4qV65c6O/Mnz/fmrskOzAkyhdTtgK6G9m6r9/GbF1yDovOpmPSlngkZOX1i5bVC880D8QrbYPhra1lICJytYDPfCpXii3LFzm3cF8P9Kzug78uZeJCik41ib+1Coswk+NKydZj2o5EfH8iv21gdX93fNE9XGWmExG5dMD3wQcf2O5IyKGMqOOnAj5tWpcBHzmq3bFZGLfhuuogoxlexxcfdg5DmA9boxGRc7Aq4AsJCVHr9kaNGmW7IyKHMKS2H6ZsTUBObl55lv+047QuORadPhczD6Rgxp4ktTxBBHq64Z2OIbivvj/X6hGRU7Eq4Hv00UfVJQM+1yMjH72q+2D1pUxcTNXh79hstIvgtC45hnPJOXhyUzy2RWcZtt1WxQtfdQ/HLcFWNyAiIrI7Vs1XhIaGqsucnLyyBeTC2bpn89c+EdmzX0+loduiGEOw5+4GPN8qCCsGVWGwR0ROy6qAr0WLFury+PHjtjoeciCDo/zgdeM/aNGZDFW7jMheJWbp1Vq9cRvjkZSd978aFeiB5QMrqyxcL4n8iIiclFUB32OPPaYu33vvPVsdDzmQUB939K6el8F4KU2HXTH502NE9mTr1Ux0XRSDX0/n1428u54fNg2PQMdIZuESkfOzarFK165dMW7cOMyePRsXL17ExIkTVSHmourwVapUyZq7JDszoq4/Vl28ka17Nh0d+OZJdiRbn4t39ibjg/3J0N8YgA72zuuYcdct/hV9eEREjhHwSZFlza5duzB27Nib/k5CQoI1d0l2ZmAtX3i7A1n6vKK109uHwN2NU2NU8U4l5uDxjdex+1p++7/Okd74onsYogKZmEFEroVFpsj6ad0aeU3kr6TpsZPTulTBcnNzMf94KrovjjEEe55uwP/dGowlt1dmsEdELsmqj7kcrSMxsq4fVl7IMBRh5pooqijXM3SYtDUBS87l/T+KesEemN09HG3ZDYaIXBhH+Mgm07o+Hnnfy7Qus3WpImy4nIEui2JMgr2xDf2xYVgEgz0icnkM+Mhqwd6SrZs3rXs1XY/tRsVsicpapi4X/9qViOGr4tSyAhHm44bveofj4y5hCNRqBxERuTCbrFy+du0avvvuO2zduhXR0dHIzs7Gtm3bDNcvWrRIFWcePnw4PD25WNpZp3VXaNO6Z9PRuSpLXVDZO5aQjcc2xOPA9fzEjJ7VffB5tzBU878x7ExERNYHfGvWrMEjjzyCpKSkQveRYFD2Cw8PR69evfiwO6Hbb0zrZuqAxWfT8Xb7EHiwkC2VYWLGN0dT8equRGTo8rZJtvhrt4XgqaYBzBQnIjJj1VyHdNh44IEHVLA3aNAgzJ071+J+I0aMUJdLliyx5u7Izqd1+97I1o1O12Mbs3WpjMSm63DPmjhM3Z4f7DUO9cRfQyMwoVkggz0iIlsHfDNnzkR6ejpGjRqFH374ASNHjrS4X6tWrdSl8TQvOee0rmbhmfyOBkS28ueFDHReGGMo9i3GNQnAuqERaBHuxQeaiKgsAr6NGzeqy0mTJhW5X40aNdTl+fPnrbk7snMDavnC1yhbV6e1NiCyUnpOLp7floDRa+IQm5GXmFHF1x2/9K2E9zqGwk8K7RERUdkEfDExMQU6blgSGBioLrOymL3pzIK83NGvZt60rrwpb2G2LtmAJGT0WhKD2UdTDdsG1PTB1hER6F8r7/+NiIjKMODz98/rRRkfH1/kfhcuXFCX7KPr/EbW4bQu2YbUc/zkYDL6LInB0YQctU1GkN/vFIKf+lZCFT9m4RIRlUvA17x5c3Up5ViKsnDhQnXZpk0ba+6OHICMuPh55E2vLT6XjhxO61IpXE7V4Y4/4/DqriTVp1nIGr31wyLwaONAuLFfMxFR+QV8o0ePVpdvv/22qr9nyV9//YX3339ffX/vvfdac3fkAKTIbf9aeTX4rsm07tX8xfVExbHkXDq6LIrG+sv5/zsTmwdizZAqaBzKxAwionKvwzdmzBj8+OOP2L59O7p374677rrLcN2nn36KLVu2YMWKFapmVr9+/TB06FBr7o4cxMg6/lh0Nr+3bo8bXTiIipKSrcfLOxMx/3iaYVs1f3dVRLkn/4eIiCpuhE+6Zvz0008YMGCAGuGTIE/zyiuvYPny5SrYGzx4MObMmcNpGBfRr6YP/G9kTUpfU07r0s38E5uF7otiTIK9YbV9sWV4BIM9IiJ76LQRGhqKn3/+WdXYk7V6hw4dQmJiosrMbdq0qWqnJqN/5DoCvNwxoKavarEWl6nH5quZfNMmRUr1bInOxpFYDzTxzka3Gt6YdSgVM/YkIedGFZ8ATze83SEE9zfw54dEIiIbsVlj206dOqkvIjGirp8K+LRpXU7JkbTcm7YjAZfTJAvDBziWBG/3/KQMcWtlL3zVPRz1Qthzm4jIbqZ0iYo7rZvNbF24erD34LrrN4K9fFqwJ/8pU1sFYeXgKgz2iIjKAAM+KhP+nu64/UZR3OuZemy6wmxdV57GlZG9ovquVPZ1x0utg+Dlzo4ZRERlgQEflZkRRkWYZVqXXNPW6KwCI3vmpDOL7EdERGWDAR+VGWmzJgvwtdpqnNZ1TdHpOpvuR0REJceAj8qMNLQfGJU3rZuQlYsNRoV0yXXsiyveyF0kW6UREZUZBnxUftO6N7J2yTXk3Fi7N+tgapH7yRhwjQAPdI70LrdjIyJyNS4R8F27dg1Xr14t1r5xcXG4cOECcnJy7GJ/R9e3hi8Cb0zrLj2XjixdUUv3yVkkZOoxanUcvjhsGuyZp2RoP89oHwIPJmwQEZUZpw34JKCaNWsWmjdvjvr166Nx48Zo2LAhPvjgA2RlFZxiWrdunaojWK9ePbRo0QK33HIL3njjDWRnZ1u8/bLe31n4erph0I1p3cSsXJP+qOScTiRmo+/SWKy7ca4l3v+ocyjm9wpXrdKMVQ/wwLxe4RhmNBJMRES2V6LqpqdPn7b6DiXQKWs6nQ5jx45Vrd20biDBwcG4dOmSCrJGjBhhchybNm3CqFGjVJAo+8nXxYsXVXAoo4Mff/yxye2X9f7OWIT5l9M3ijCfTUf/G+VayPn8dSkDD6+/jqSsvJHcSj7umN87HF2q+qifB0f5Yv2FFBy5FIsmNaqgZ61AjuwREdlbwNe2bVur7zAhIQFlTXr6SrAXERGBzz//HL1791YtmuLj49V13t6ma4VefPFFFYxNnjxZ9QD28vLCzp07cdddd2H+/Pl45JFH0Lp163Lb39n0ru6LYC83JGXnYtn5dGTqQuHjwXprzkR6Zn9+OBWv7kqEVmO7aZgnfuxTCbWD8l9mZNq2S6QXorJ0qBXpxWCPiMgep3TDw8Ot/iprMl0rU7li3rx56NOnj6EfZ1hYGF599VXUrFnTsP/Bgwdx+PBhNeX72muvqWBMtG/fHi+88IL6/qeffiq3/Z11WlfL1pWRn3WXMyr6kMiGMnW5mLglAS/vzA/2ZBp/1eAqJsEeERG50JRuWdu1axdiY2Nx2223qTVzMrImU7kyrRsSElJg/7///ltdDho0qECj9qFDh6oAUdunPPZ3ViPr+uHnU3nTugvPpOP2Wlyz5Qxi03V4YO11bI/JXxc7tWUQXm4bBHez/3ciIqo4Tvfx+9ChQ+qyW7duarr03//+N65fv66CrVtvvRUzZsxAu3btDPufO3dOXTZo0KDAbdWuXRt+fn44e/Zsue1fmIwMxx4V6xQOo2ndDCSmptt8WldLxrGUlEO2dyg+B2M3JuPSjS4avh7AzPaBGFnHG1mZhSfn8DzZP54jx8Dz5NrnyNfX17UDPlmnJ/bv34+ZM2fCx8dHTeFKWRYZSZNRtZUrVxrWzKWkpKhLSaSwRLYnJiYafi7r/Qtz+fJllYziyLqFeWNZjCeSs3Px24Er6F6pbP6e6OjoMrldyrf2mgdeO+6NDH1e0F7FW4/3mmShmUcaLlwo3iPF82T/eI4cA8+T650jDw+PEifBOl3A5+6etyzxr7/+wjPPPKOSJCTok2SRKVOm4I8//sCbb76J3377zWR/vd5yr0+ZEvb0zH+Yynr/wlSvXh2O7h73LCyLSVbfb0sPwphaQTa9ffkEJU+qyMjIAok5ZLvkjJmH0vHu0fwi2m0qeWJu1yBUNSu5wvPkuPhccgw8T/Yvy47el2wS8MnI2YIFC7B3715VZiSziOmc48ePoyxJYoaQyPf11183rJuTNXySobtixQps3rzZsL9sF3LcloIxCRSrVatWbvvbaujWHg2o44Pg7SkqcWPVpWzA00cldNiaPKmc4fGyN2k5eozflICFRh1TRt/ih4+6hKk2eiXF82T/eI4cA8+T/fO2g/clqwM+Capk6tReSHFlUatWrQJJEv7+/qhSpYrqdCFRt5wAbf99+/YVuC3JsJVpVOP1d2W9vzPz9nDDkCg//HAyTU3rSs22wbWZvOEILqbk4L6/rmP/9bxC4fLMeu3WYExqEVjgeUZERE7WaWPJkiUq2JNpS5kuXb16teG6bdu24fvvv1flR8T48eNV7bmyJvcXGBiIAwcOFKj5d+rUKVXw2HhoVZI75PgXLVpUYP85c+aoyx49ehi2lfX+rpCtqzEeKSL7tSsmC72XxhqCPWmV90OfcExuGcRgj4jIFQK+uXPnqsuJEyeqGnPG2a9NmjTBkCFDsGzZMlWwWQogx8TEoKzJkKl02ZDM3JEjR6r7l9G1X375BXfeeadagySXGinOLIkckuwhhZClBZpMTUt2r2T5SvB43333ldv+zq5HNR+EeueNCK04n4H0HPbWtWc/nkzD4BWxiEnPW4NaO9ADfw6pgoFRHJklInIkbgkJCaV+x5V1chJYycidNnWprVmTAEeb6lm/fr1qZzZw4ED8+OOPKGuSGTt8+HDs3r27wHUSfC5evFgFWhrJ4O3fvz/Onz9vsq+MzEmgevfdd5tsL+v9nd3Tm+Px/Yk09f13vcMx1EbTulK6RqbrZTq/otdKODqdPhf/3p2EWQfzssxF16requ9tJam/YgWeJ/vHc+QYeJ7sX4YdvS9ZtYYvOTkv41L+EI10ksjOzkZaWhoCAgLUtlatWqnL8iowLMGctFabPXs2Vq1ahbi4ODXSNmDAADz88MMFHvSqVati48aNqkOHJHTIsUsA+8QTTximpMtzf1eY1tUCPinCbKuAj2wjMUuPcRuu48+L+clXjzQKwDsdQ+DlzvV6REQuN8LXvHlztSZOMm8loBISyMjUrXS80JIRZO1anTp1VDAoXTDItWXrc9HwpyuIz8xFgKcbTtxbFf6eVq0usLtPUo7qdFIO7l0Th2OJOepnqY39TocQPNYkf0TcWjxP9o/nyDHwPNm/DDt6X7LqXbZu3brqUv4Y4yBQrFmzxrBN1q2JSpUqWXN35CRklEgb1UvNycVqo5EkqjgbLmeg95IYQ7Anay3/17+yTYM9IiJywIDv9ttvV5d//vmnYZskSmjlWqZPn44PP/wQkydPVtv69etn3dGS0xhZxyhb9wyzdSuSJDLNPpKCO/6MQ0JW3oB/oxBPrB0agR7VfSr02IiIyA7W8Ekixq+//oojR44YtknG6e+//64SNd59913DdhnOlK4XRKJbNR+E+7jjeqYeqy5mqKK+tpjWpZJPr7+wPQFzj+WtqRT9a/pgdo9whHjzfBAROQurAr4aNWqowM68v5u0LZPOGxs2bFBdNyRpY9y4cYYuGESe7m4YVtsX3x5PQ9qNad3hRqN+VPbiMnQYu+46tlzNb+r9TPNAVVDZg8kZREROpUx66Upv2AcffFB9ERWVrSsBn/jjTDoDvnJ0OD5bJWecS9Gpn2UwT1qk3VvfvzwPg4iIHCHgk1E8MWbMGJvuS66hS1UfVPZ1x7UMPVZdyEBqth4BXpxGLGvLz6fj8Q3xSLlR9DrCzx0LeldCu4iKbexNRERlx6p31wkTJqgvW+9LrjOtO7R2Xpp6ui4Xf17MqOhDcvrkjA/2J2PMX9cNwV6rSl5YO6QKgz0iIifH4RSqUCPq5E8hyrQulQ1pYff4xni8sTsJuUaZ0isGVUbNwDJZ2UFERHakXF7pr127pi79/bk+iEx1qeqNKr7uiM3QqxG+lGw9Ajmta1NX0nQY81cc/rmWbdj2SpsgTG0VZGh/SEREzq1EAd+5c+dKtF1rvzZnzhz1fb169Up6fOQK2bp1/PDN0VRk6KDW8t15Cz8Y2Mo/sVkYszYOV9L06md/Tzd80S1MPeZEROQ6ShTwaT1xi7vdHBM2yJIRNwI+bVqXAZ9t/HoqDRO3xKtAWtQM8MCPfSuhRbgX/xGJiFxMiQK+gIAAk59TU1MtbtfIdJGfnx/q16+P0aNH46GHHrLmWMlJdY70VpmiMel6rL6UgeRsPYI4rVtq+txcvPlPEj7Yn2LY1inSG/N7haOKn4dtThoRETlvwHfp0iWTn0NDQy1uJyoJKfI7vLYfZh9NRaYOWHk+A6PqcVq3NCRYfmJjPJafz894vr+BP97vFAofD67XIyJyVVYlbbz22mu2OxJyaSPq5gV84o+z6Qz4SuFscg7uWxOHwwk56mdplvFWuxA82TSAyRlERC7OqoBvypQptjsScmkdI7xR1c8dV9P1WHMxA0lZegSzl2uxbb6aibFrr6vexCLY2w1ze4ajT428OodEROTabFaWZc+ePfjzzz9x4sQJlZkbFBSEBg0aYMCAAWjdurWt7oaceFpXMke/OpKKLD2w4kIG7ua0brF8eywVU7cl4EYtZdQP9sSPfcPRIITJGUREZKOALzY2FuPHj8fq1astXj9jxgwV9H366aeoXLmytXdHTt5bVwI+LVuXAV/RcvS5eGlnImbfeMxE7+o+mNMzHKE+rKlOREQ2CvjS0tIwfPhwHD582FCepU2bNoiMjMTVq1fVqN/+/fuxatUqjBw5Uo0AStYukSUdIrxRzd9d1YxbeykDiVl6hHBa16L4TD0eWncdG65kGrY91TQA/2kXomobEhER2Szg++KLL1SwFxISgi+//BK33357gX2WL1+OJ598EgcOHFD7TJ482Zq7JCfm7uaG4XX88MXhG9O65zNwT31m65o7lpCNe9fE4XRyXoE9qWAjWbhjG1ouj0RERGTVvM8ff/xhmLa1FOyJQYMGYfr06Sb7ExVG+rsa/r/OunZvXZ0+F5uuZOK302nqUn5efTED/ZbGGoK9yr7uWHx7ZQZ7RERUdiN8p06dUpeDBw8ucr8hQ4bg6aefNuxPVJh2Ed6o4e+BS2k6Na2bkKl3yfVoi8+mY9qOBFy+0RJNBHu5ITk7FzdyM9AsTJIzKiEqsFxaYhMRkQOz6p00NzfvrcfTs+g3HO16vT7/zYvI4j+km2Tr5pUSydYDy8+nu2Sw9+C66ybBnkgyCvaGRPli1eAqDPaIiKjsA76oqCh1+ddffxW537p169Rl7dq1rbk7cqFsXc1CF5vWlWlbGdnTAjtLgrykxl4YAtl+joiIyiPg06Zyp02bhn379lnc5+DBg+p6bT0f0c3cVsUbNQPyer6uvZSppnVdxdborAIje+ZkWnd7THa5HRMRETk+qxb/TJw4ET///LPqpduzZ0+VuNG2bVtVliU6Ohp79+5VWboy9VuzZk21P1Fxs3U/PZSiigkvPZ+O+xu4RgZqdLrOpvsRERFZHfCFhYVh0aJFePDBB3Ho0CGsWLFCfZlr3rw55s+fj9DQUD7qVOxpXQn4xMIzrhPwSTHl4oj0yxsBJSIiKg6r0/vq16+PjRs3YtmyZarbxsmTJw2t1eS6/v37q6lcDw++QVHx3VrZS03rXkzVYf3lTFzP0CHc13n/h2QU/LsTaXhxe0KR+0lJ5eoBHugc6V1ux0ZERI7PJvUcJJgbNmyY+iKyBTc3NzXKN+ugNq2b4bS15qRrxqQt8Vh8LqPI/bT+GTPah6jew0REROWStLFgwQL1Zet9icyLMMu0rjPaeCUTXRZGmwR7Yxv6Y3b3UFT3N316ysjevF7hGGb0uBAREZX5CN+ECRPU5ZgxY2y6L5FoU9kLUYEeOJ+iUz1j4zJ0qOQk07pZulxM35OEjw6kGEqwhHq74eMuYYaA7o66/iprVxI0ZM2eTONyZI+IiEqDJfrJvqd16/jho4Mp0Mm07rkMPNjI8ad1TyZm47EN8dgbl19apVtVb3zRPRw1bpSjERLcdavmU0FHSUREzqRcelZdu3ZNXfr7+5fH3ZETcaYizJKYMf94KrovjjUEe55uwOu3BWPhgMomwR4REVGFjfCdO3euRNuFZOzOmTNHfV+vXr2SHh+5uFaVvFAnyANnk3Vqvdu1DB0qO+C0rqXEjPrBnvi6RxhaV2bGLRER2VHA16pVqxJtN8f1e1Taad2ZB/KndR9ysGldCVSf3GjaG1cSMyTbNoDt0YiIyN4CvoAA0zfa1NRUi9uN36z9/PxUPb7Ro0fjoYcesuZYyUWNqJsX8Ik/zqQ7TMAniRkz9iThQ7PEjI+6hKlOIkRERHYZ8EkLNWNa5wzz7US21DLcC3WDPHAmWYdNVzMRm65DFTvvNCGJGeM2xmPPtaITM4iIiOw+S/e1116z3ZEQ3aQI8wf7UyCdx5acy8AjjQPsumPGtB2JSJOK0TcSM15tG4yJzQNZVoWIiBwv4JsyZYrtjoSoCCPq5AV84o8zaXYZ8DExg4iIXLosC5G1WoR7oV5w3lToFilGnKazqwd105VMdF0YU6BjxoZhVZiFS0REFY4BHzlQtm5eHce8ad10u0nMeP3vRAxbeQ2XbgShkpghLdCkawazcImIyB4w4COHytbV/GEHRZhPJeZgwPJYlUGca5SYsWVEJLNwiYjIrjDgI4fRLMwTDULylp1uvZqFqxU0rasSM1THjBhDFq4kZvz7VnbMICIi+8SAjxxqWleSN4SMqC2ugFE+Scx4aP11TNySgNQbWbiytnD1kCqY3DKIWbhERGSXGPCRw/bWLe9pXS0xY9FZ88SMCLRhezQiInLWsixE5a1JqCcahXjiWGIOtkdn4UqaDtX8y7aQcbY+F9P/YccMIiJyXBzhI8eb1q1bftO6kpjRfxkTM4iIyLEx4COHo63jEwvLKOBjYgYRETkTTumSw2kS5qWmdo8k5GBbdBYup+pQ3Yb9aRMy9Zi0Nd5krZ4kZnzdI5xr9YiIyCFxhI8cvibfIhuO8kliRhezxIwHGjAxg4iIHBsDPnJItp7WlcSMwjpmzOoahkAvPlWIiMhxcUqXHFKjUC80DfXE4YQc7IjJwsWUHFT2LH1ixmMbrxuKKGsdM77oHo4aNpwqJiIiqigctiDnmNY9lz8FW1xMzCAiIlfBgI+cY1r3TFqJflcSM9gxg4iIXAWndMlhNQz1Uv11D8XnYFdsNi6kFq+37uarmXhiQ7xhrZ6WmDGjQwjX6hERkVPiCB85tJF1/Q3fLz2fddPEjDd2J2LoCiZmEBGRa2HARw5tRB1fw/eLL2TdtGPGB/tTVIcOLTFjy4hIDDeaGiYiInJGDPjIodUP8UKLcC/1/Z64HFzOcDO5nokZREREXMNHTmBkXT8cuJ5XUuWzs17ICc5Gz1o+SM7OxeStCSZ1+tgxg4iIXBGTNsjhBXnlj+qtuuaJVWuTUNk3BXo9cD1Lb7iOiRlEROSqGPCRQ1t8Nh0vbE8ssP1aRn6gJx0zPuoSxrV6RETkshjwkcPS6XMxbUeCIQnDEm93YMOwKqgdlLfOj4iIyBUxaYMc1tboLFxOyx/Js0RmdM+nFL0PERGRs2PARw4rOl1n0/2IiIicFQM+cliRfh423Y+IiMhZuUTA99VXXyEqKkp97d+/3+I+Fy9exMSJE9GmTRs0btwYw4YNw/Llywu9zbLen26uc6Q3qvu7w7TyXj7ZXiPAQ+1HRETkypw+aUMCrTfeeAM5OTnIyMiATldweu/MmTPo27cv4uLiDNuuXr2KjRs3YsaMGXjqqafKdX8qHg93N7zdIRQPrruugjvj5A0tCJzRPkTtR0RE5MqcfoRv6tSpuOWWW9C/f/9C93n++edVMCZB2ebNm3HkyBFMnz4dHh4e+Pe//40LFy6U6/5UfMPq+GFer3BU8zf9V64e4KG2y/VERESuzqkDvkWLFmHNmjX4+OOP4enpWegIoOwTGRmJ+fPno3nz5qhWrRrGjx+PJ598EpmZmfjhhx/KbX8qOQnqDoyqit97B+PNRpnqcv9dkQz2iIiInD3gS0xMxIsvvogJEyagdevWhe63ZcsWdSlr6vz9/U2uu++++9Tlpk2bym1/Kh2Ztu0S6YUBVXTqktO4RERELhDwyVSpn58fpk2bVuR+J0+eVJctWrQocJ0kV3h5eeHUqVPltj8RERGRrTll0sb27dsxb948/PHHHyrou9lIoAgPDy9wnayxCwkJQVJSUrntXxhJOKGiZWVlmVySfeJ5sn88R46B58m1z5Gvr69rB3zZ2dmYPHkyxowZgx49ehT799zcLGdyuru7Q6/Xl/v+5i5fvmwxw5gKio6O5sPiAHie7B/PkWPgeXK9c+Th4aESUl064Js5cybi4+Pxn//8p1j7BwUFqcuEhIQC1+Xm5qoROhmFK6/9C1O9evVi/T2uTIL92NhYVKlSRU2Vk33iebJ/PEeOgefJ/mXb0fuS0wV8s2fPVsFVy5YtTbanp6ery6FDh6pRtdWrV6NRo0aoU6eO2n706NECt3X69GmVRWscRZf1/rYaunVF8hhpATbZL54n+8dz5Bh4nuyfrx29Lzld0oZMe0pELevijL9km0hJSVE/a9OjnTp1UpdLlixRxZmN/f777+qyQ4cOhm1lvT8RERGRrTldwLd3716cPXu2wJeM7GmBl/wsGbKifv36aNeundo2adIkNcUqU61Lly5V08MyGnjvvfcabr+s9yciIiKyNbeEhATjjlRO65FHHsH//vc/rFu3TvWzNQ8SBw4cqKZ9JbnC29tbTbUKCdJef/31ct2fiIiIyJacboSvNKQw84oVK9C9e3eV+SLBWFRUlOpzK/X8ynt/IiIiIltymRG+tLQ0VQdHFk9K0FUYKZEi6/18fHyKdbtlvT8RERGRtVwm4CMiIiJyVU5XloWcz4ULF1SSy9WrVxEREYHBgwcbyt2UhJTBkal1uT35/SeffLJMjtcVyaj1qlWr1HpVWad66623ol+/fkWOplu6DekrffjwYcTExKi6Vd26dSuyFzaVjLR6lOfAtWvXUKNGDQwZMqTENT6l7NXatWvVbcm6ZHku3X777YiMjOTpsAHpqrR8+XIcOnRI1W3r2LEjevbsadVt/vTTT+q5KTVfX3rpJZ4nGzh48CDWrFmjng+1a9fGsGHDUKlSpWL//o8//oh9+/ZZvK5JkyZ48MEHYWsc4SO7tmDBAjz77LOGJBchL4LvvvsuHn744WLdxoYNG/DCCy/g2LFjhm3yIrpy5coyOWZXI0VF77zzTuzfv99ke/v27fHrr78Wq7D4d999h9deew3Xr18vcJ0kPH3zzTfw9/e36XG7mlmzZqk1w8Yde+Qx/fzzzzF8+PBi3cYbb7yBzz77rECrR1miMn36dDz66KM2P25XItUc5Llk3l+9f//+mD9/fqnqsUoNWFk/LkuaJLiXD1RkHXmt+vjjj1XFDU1oaKhq6VrcDl9aIqklAwYMwM8//wxb4wgf2S359PPMM8+oNyh5Q5LsavnU+9tvv+G5555D8+bNVcmbm5FAT77q1q2LLl264Pvvvy+X43cVTzzxhAr25FPu3XffrdapyojCzp07VRb6t99+e9PbkN9PTk5Gnz59VEH0ypUr48iRI6oftoxIyajERx99VC5/jzOSEbl//etfqgzUPffco8pS7dq1C8uWLcPjjz+unkv16tW76e3s2LFDBfAymiH7S23RrVu3YsuWLYbnJOuKlo48bx544AEV7Mn5ueOOO9Tac3m9+vPPP/F///d/6oNuSUhAIq1G5ZzI85GsJyNz8lok1TakhaskYMqgwvr169X5++eff9TrV3HJeTUP5Eszg1UcHOEju6V9AnrxxRdNpiFkpELevGQaSQKLmzl+/Lh64ZNAQvoZyiVH+Gxjz5496NWrF2rWrKmmY8PCwtR2mZKV4FpG/ySwaNCgwU2De3nh1H5fI1Mmd911FwIDA3Hx4kUbHbXrkWUQEpR98MEH6nmlefXVV/HJJ5+o6aPiBNRyvqWLkflUvYygf/XVV3jqqadU9QEqOZlxkGC8adOm+Ouvv+Dn52dYiiJLG2SETj64hoeHF/s258yZg1deeUWde2kCIFOOHOGzjgwynDhxQgXisiRCIx+cfvnlFzz//PPqMS/u+5uM6sroYHlgWRay20+70v5OPvnIKJ/5iJKMMkhNReOp3sI0bNhQBXlkezL6Jh577DGTYE3WWj700EPq++JMnbdq1apAsCf69u2r1pqlpqYW6FRDxSPF3rdt26bW2JmvC5JROQneiru8QUbZLa3LlPMkJCih0tHOwdNPP20I9oS03pQPPbLGVT4AFZeseZYpfPnAXJz2nXRzMvoqwZ4E5cbBnvahR9jzUiEGfGSXzp8/r6b4ZDQhICDA5DoZSpdPWRLsycJxqjjaaIHWQtBY165d1aVMw5eWtELUemN7enIFSmnIGi75ACXPGfNgTYJsefOSkW9J5CitxYsXq0sZ1SXrnkudO3cucJ32uJbkuSSBnkwNSgBJtqE9/pZe76SrlnyoklFY43WyN7No0SLVfEFGxiU5sSw/NPEVlOySTAkKGd2xRNsuU4ZUcbTH39J5ssU5evPNN9XoHjMLy/a5dODAAbVfcdceffHFF2oqSvqS//3332rZxH333YeRI0dacaSurajzpG3T9rkZGWWS4EGmhvlBqXxe77Tt8uFJPqQWN2NX1jkbk+UxMhUvSW+2xoCP7JKWBSijeZZohavNswWpfGmPv6VC4taeI8nMlcBCXhBlvSaVjrbsobBi79qC8eIsj9AsXLgQ27dvN8mklqBcSvJQ6c+TPH5ShcCacySj4lOnTsWECRNY0sjO35caN26sZkIke/rKlStYsmSJWqssU/iSZFO1alUbHj0DPrJT2hqWwp44Uv/LeD+qGNrjr50PY9q20pRT+fTTT1VCwbhx49hr2kpasGDpHJX2uSQ1LCVTV9YHSj0yySKVNy7Jqm7btq21h+yy50mSyySoMw/OJVu3uOfoP//5jwoaOSpece9L/sV4zZMpd/O15TKjIZUOJON37ty5Nj+HHOEju1StWjXDWj5Lzp07Z7IfVfx5kixbS+eopJ9S33rrLbz33nsqOeedd96x4dG6Jq2w8s2eSyU5TyNGjDD5Wd6gZJuUmJCpRCrdc+nMmTPqPJlntRf39U6SCmbPnq3WmMm6MGOS9CRT8NOmTVNT9zIKSLZ9X5LtEhQWp/aopURCCfSl7qw8n6xZ+1wYBnxkl2Qdg5QfkLVFUozXuBSBTFns3r1blepg9lnFatGihRrVkTpUWpKGRl60tH2KQxILpKSBTOXKdJQEfmQ9eWORER+poWc+enT58mW1yLxWrVpWlYaQwr4yQiUJIlQ68jyRmobyXDIP+GSbtk9RZI2fPI+kDIt8mZPXTlkmITVJGfCVnPb4SwkqGY01XsIgpaXkveq2225T9S5LS0vaKIu1l8zSJbs1aNAg9alUqvsbk2wmWcgv67q4ILniz5H4+uuvDaMQQkoXSGcAyQrV9imKlJyQ6VsJ9qRQLIM925Esd6mVKAvJ//vf/xq2yxuWlO2Qy6FDh970diSLV0ohWSLTTzLNxRF362olCqmHaJwxLUkxsmZSzmPv3r2LvA3JypWOJ5a+5LUyODhYfc9gr/QDEVJCSkZi5bXK+PVLe58qznNJ2ntKcG8pKUS7nbJoKcnCy2S3pOCojBrJ+hVZFyRPACldIIvFZZRi48aNJsPi8mYk/VxlAblxext5Er3//vuGNRbS/kbemLR2UjLlZV7rj4pPavBJ9xOZxpDHXkYYpD6flNWR64yDDAngZW2ejM7KpfF6li+//FL1z5XWUpbIPpZq9dHNyYi4tGuSx1/KfshicdkmoxISBJgvEJdSEVK7b9SoUaovspBMXMkclNEh6fUpzyGZIpRizFp5JCnifP/99/OUlJKsi5TXNXkeyPmS1z7pqyvB9Msvv2yo9Sbi4+PVkgc5b/Ih6WakZAgLL1tP3mNknZ2QzkDSYUhGU2WkXN5LZCQ9KCjIpD2ozFRJoWWpCSvkuSWvlTKSK89FOd+XLl1S517eo+ScyvucrQsyM+AjuyaFRmUtV1xcnGGbBBYyLSFPGGOy7ktGhqTP4ZQpUwzbtTeqwkg7qM2bN5fRX+D8ZLRVXszkhdCYrOmSIM54ClHeuOTFTKboJaDXjB071lDLrTASnMiLK5XO77//rgIDCcSNgwAZnTOv/SZT67IWTJJnpH2UFmBIJw0piG5eZ0wCcQng2UvXOjKyJ+25JCAwJs8v+eBkPFUoI+oy2iTBt/n+ljDgsx15bsj/u3HWtLw2SfcN82l37bVNPhRrBcolE1eWrUiAZ9yPV0i9THneacGhLTHgI4cIKGT0TuobySchmZ4y/gSlkREL+XQlC5alI4BGprJ++OGHQm9fFjCPHj26zI7fVUhAJl9CRoWaNWtWYB8JFCQIlIXNDz/8sGG7BIvmDePNyciRjEZR6clzQdaDSWAhoxE9e/a0mPkpa5RkVEKmEGUEwpg8D/fu3avetGRtoPTUlTepwkpVUMlJO0JZtC+Pr7SBtNTnWEZXJcCQUTttxKko0vpO1llKAELWk+eQrFOWTHUJ9mRWyVJJHe21TTpzmCe2SZKH9BGXrijyniYBvPnzzZYY8BERERE5OSZtEBERETk5BnxERERETo4BHxEREZGTY8BHRERE5OQY8BERERE5OQZ8RERERE6OAR8RERGRk2PAR0REROTkGPAREREROTkGfERUYh988AEmTpyIgwcPVtijJ30s5RjkS6/XV/jtlLWUlBTDcRbm+vXr6nrphevI5s+fr/4OaV1VXv73v/+p+1y+fHmJfk+eA/J78pwgsmcM+IioxORN8bvvvlP9VCtKdna2Ogb5siZQs9XtlDUJTLXjLExycrK6vqje0Y5g8+bN6u84fPhwud3nzp071X1q/aCLS54D8nslDRSJyhsDPiIiIiIn51nRB0BEVBq+vr74+OOP1fceHh4VfjtERPaMAR8RWZSbm4s1a9ao6bW0tDTUrl0bw4YNQ1RUVLEesZiYGKxcuRLHjx9X05HVq1dH37590aJFiyJ/79q1a/jzzz9x9OhRpKeno1atWmjXrh06depksp9Op8OuXbvU9/fffz/c3NxMrv/777+xYcMGXLlyBX5+fqhZsyY6duyIVq1aleh2tHVa8lhcunQJ3t7eaNKkCQYNGoTw8HCL6+hee+01FUi+9957SExMxJIlS9T0pNxX06ZNcccddyAoKAjF9fPPP+Ovv/4y/Gy+ju+ZZ55BgwYNCvxeSe7b/LgTEhKwdOlSHDt2TN3Os88+izp16hj+N+T/YuvWrYiNjYW/vz9atmypHhP53hK5PZn2lGPJyMhAZGQk6tWrhwEDBiAgIKDIv1/OpTz+cl+VKlXC7bffjrZt2950zaP8/+3fv1/9/0ZERKBnz55o3749SkPO/R9//IFz586px09up1+/fqW6LaKKwICPiAqQN/gxY8aoN3Vjb7zxBmbMmHHTR0wWsEvQIAGbsddffx2jR4/GRx99pIIwYxJEvP/+++rL/PdEs2bNsGXLlgJr78TMmTPh7p63QkV+99FHHy10TZUEj6tXr77p7Wi3JcHUr7/+WuB2goOD8e677+Kee+6xuI5Ogpjhw4fjkUceUcGvsXfeeQcLFy5Ew4YNURzbt2/HL7/8YvjZfB2fPKbmAZ+cu5Lct/FxS0D12GOPIT4+3nC9/D9IwHfixAn1+EogZU6CuK+//hrdunUz2S7n4sknn0RSUpLFx/HTTz/F0KFDC1yXk5ODcePGFXj85XGfNGmS+n+yRM7vhAkTCvzt06dPR58+ffDNN98gNDQUxfXTTz+pgFcCR2O33norHn744WLfDlFFYsBHRAXIm5gEDF5eXrj77rvVqJiMvP3444+YOnVqkaNTEhRKwCcjZTJ6c9ttt6kgQkbsZKRKC1y++uork9975ZVX8Nlnn6nvZfRGfldG0C5cuKBG4LZt21asMyUBowQYMlI1cuRINarl6emJ8+fPY8eOHThw4ECxz7gEGzLKJX/LiBEj1Ahhamoqfv/9dxw6dAhPPfUUAgMDMWTIkAK/K6Oa9957rxoRfOKJJ1C3bl01OiQZqJcvX8b48ePVqFVxyDmQgO7ll19WP2tT0BrzYM+a+5bfffDBB9U5k6BKgjyZ6pbbkHMhwWBcXJwasZXHV0Zg5ecVK1aokVA5VhmNlFFQIdfJ4yiPW/PmzdUooIy2yf/TqVOn1O+dPXvW4rHI/5GMDEqQJiO8ch5l1E4CYPnQ0KtXLzVqZ0z+V+Rvl2BRjl2Op3LlyipAlcBNjm3UqFHqfuX2bmbTpk3q8ZKEHvlfuvPOO9U5l//HRYsW4eTJkze9DSJ7wICPiEzIdOratWtVkPPbb7+hR48ehuuefvppDBw4sNByLBJMySiZj4+P+l3zkR5545RpMAn65PvWrVur7f/8848h2JNpxSlTphS47eKWgJGAQMjtyfSlpWMsDgkMJNgTMiJkfFsy6vfAAw+ooOHFF19UQZB58CABhwRDixcvVgGHRgIGeQxkmlKmu4szyieBpnHAN3bs2CL3t+a+5XerVKmi/n7zKWsJpCSAk9E4GcmT86x56aWX1FTzggUL8Oabb6pLIR8cJNiTKV+5TfkQYT6yKLdpiQR7Etwa/70ShN51113qtiQb2Tzge+GFF9TfIB805O83nmKW45djl6BQfvdmj6P2QUSCPQk6JWDUjl8C6e+//149J4gcAbN0iciENn0mb4zGwZ6Qkb3CptHEvHnz1NSsjOiYB3uicePGeOihh9T3WjAlvv32W3XZvXt3i8GekNGhYr2o3ZiSveWWWyxef7M1hBoZjdSOyTxwlOBOprblzV/Wdm3cuNHibcg+xgGXkEBEgh9RlmVHrLnvf/3rXwWCPRkZlGBaRg0/+eQTk2BPe9zfeust9UFBgjEZKRTamkhZ+2ke7Gn/U9raQHNdu3YtEJTJ7Wn/QzLKakyC2D179qjv5VjM1xN27txZTX8LGa2+GVm/KCOD2t9mfvyy5vNmawmJ7AVH+IjIhPaGOXjwYIuPjIx0yHSfjNqY06Zd5U2ysALB2hTYmTNnTGqgCRm5sZYEmnL/smZMRnt69+5dovVa5o+DrMOzRIKUNm3aqGOXEUq5H/PARAIsSyT4kXpvslayLFh73xJomZO/U4J5CaJeffXVQn9XAkFJypAAUaaB5ThkveayZctUIClTrLIe01JyjLkOHToU+jcI879BO2fVqlUr9HdlGlqCvb179970/rXbk+QS+bBiiUzny/knsncM+IjIhDa9VlQ2rkwXypo8c1evXlWXhY14mWdRarTF9ZIJbC2ZYpU3askglaQFCSzkzVqCGBmpM8/2LYysMbvZ4yDHK4GQtq8xCYwKyz7VRookc7YsWHPfErDJGjtzku2sTbMWVfzZ/PzKWj+Z5pekh1mzZqkvCcAly1WytiUADAkJsXgbkpFb1N9gXii7uOdMS8iRY5T1eNY8F4qbtU5U0RjwEZHFKdGsrKxCHxltus6cNmoj6/MKGxHRSJkUjVb/zhYBkGR9StKGrB1btWqVCshkNOfIkSOYPXu2CvpkTd7NRpi0YyrO4+BM9fsKS2TQHq/69eurNYw3Y3x+JZNZgjvJDpbzImsIZa2ofL399ttq+lyyp61VknNmvP/NnguF/b/f7Doie8KAj4hMyHSYjGxI5qT5Gj4hC+Jl3ZolNWrUUKMschvFWRBvHBxIjTUJysynRktLRvS0qUmZYpTgTxb8S8/U/v37FyinYq5q1arqbzl9+nSh+2jXyd/r7OTcCnlMJGGlOFOyxmQ9oZR6kS9tal/WxUltO8l2liDQWtp5kP9dmX62dIzaUgIZZTQvDWRORie12ytMUdcR2RMmbRBRgYxQIRmIlkh2Y2EjKFrGpGRoSpBVXFpgKVOFUhfP1qREi6zFk0X2Qiu0XBRt6lf+Fks9diVA0TKHtcesLBknDJTFY3QzXbp0UccgU7qSgW0tGSmUqV4t+DOu+VdaMk0sQZ7clmToWjJ37lx1WdgaP2My6ii3J/1ypYi3OTkPkrlL5AgY8BGRCa3bhARFko1pnrUoxWsL8/jjj6tRE1nfJ6M2EhyYk+K78iYpdeE0UshXgjL5PSl3YV6gVwreSgZwcXzxxReq5p45Kf8hNdW0ad+bkRFKeRzkmKTMiHHQFx0djcmTJ6vvJeu1sAQJW5JsVi3r1Dw7tTxI1q42avvcc89h3bp1FveT6XOpU6iR9ZyStSsjw+akrI2QQFLOvy1G+KREjpAEEfPRWan9uH79esP/3M1o3WG0taHaOkZt+YHch9QmJHIEnNIlIhNSG0+m3WS9m2RjyoielESR1lvy5i2L+uWN1fjNz3jaT+rfSVkWmaqTNVoy6iIdGCT4k9+RAErWPUk3BG0BvSSBSJFdqWkmU65SFFhG2GTaTd5QZSRNgkApCHwzUsdv2rRpqsac3K5MJcp9Swax3IYEpNI14makfIskGkghZzk2+XtkxEcW+svjIJeSGCFJCOVBgk8ZZZPHTYpAy6ioFrgW1lrN1qSotiTESFaqZLvKYyyPk6x1k6xcmd6U6X4pmi01/4QUu5apW0nMaNSokfofkf2lhIpWE1E+ZNxserW4pJOIfFiRoF/+hyRrW5I/5L60UjQynS/T+sUtbyN/g/zfSmAvZXok+Ja1ofL3yuMu3UeI7B0DPiKy+KYpIy4SvMmbpPZGKWvtvvzySxVUWQr4hAQCEhDK6Ie88ZqPBMkbv+xjnpF73333qYK/UlxY3kBlzZ1GAqviBHtCRhYlOJOAQr6MSTkQmUaU6cTikL9BggV5PGTtl3EpGek+It0ezHvzliUZaZRRVglmpMtDUa3VyoKcBymvIkHQnDlzLD7GUpdOC/aEBEgSHMmUqFZ+RyNBnny4+L//+z+bHaNkzWrrNSVBxLijiNyftFyTItHFJf8rkmwio9cy9ayNSkrQKh1pZM2prGkksnduCQkJuRV9EERkn6RcirSxkvV4Mlomo3WS2ShZsLJ4X97sjLMxzUlgIiMrMrImU4ISCErLLUsFeI3JlKUEfbJGSu5XWlqZT8PKFKGMPgpLSQTy5iwjMPI3SKAi91tYZ4mibkfIiKQEKzKKJYWHJQNZax1mTkb+ZJRSsl0liLVERgjl2GTtX3H76Rofr6wflMdWWycpo1WSZGLNfRfnd80fk927d6vHRB4zObdS7FqOwxL5H5BRMlkPJ+dV9pfg0FJZFJl6l+BaRlQtPc4yYqt10SiqdqPchtRklCUBMsos/7+FlWGRDyeSNCRT9FoHGGMypa+N6sltyGif/K3y90hnGhlJlrZxRPaKAR8RERGRk2PSBhEREZGTY8BHRERE5OQY8BERERE5OQZ8RERERE6OAR8RERGRk2PAR0REROTkGPAREREROTkGfEREREROjgEfERERkZNjwEdERETk5BjwERERETk5BnxERERETo4BHxEREZGTY8BHREREBOf2/yAyMlELq1J6AAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 600x350 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(figsize=(6, 3.5))\n",
    "ax.plot(thresholds['threshold'], thresholds['error cost ($)'], marker='o', lw=2)\n",
    "ax.set_xlabel('decision threshold')\n",
    "ax.set_ylabel('total error cost ($)')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c031",
   "metadata": {},
   "source": [
    "At the default threshold of 0.5, the errors cost \\$10,728, more than the \\$9,381 of flagging every session (Section {doc}`1502-evaluating-classifiers`), because only 39% of the buyers are caught. Lowering the threshold to 0.1 flags 1,154 sessions instead of 298, raises recall from 0.39 to 0.87, and cuts the error cost to \\$4,221, a saving of more than \\$6,500 on 3,699 sessions, even though accuracy falls from 0.885 to 0.802. The retailer sends more nudges that turn out to be unnecessary, but each one is cheap compared with a lost sale. Going too far does not help: at 0.05 almost every session is flagged (3,279 of 3,699) and the cost rises again to \\$8,187. Notice also that the most accurate threshold (0.4, with accuracy 0.890) costs \\$9,426, more than twice the cost at 0.1.\n",
    "\n",
    "There is a simple rule of thumb behind this. Flag a session when the expected cost of *not* flagging is larger than the cost of flagging:\n",
    "\n",
    "$$30 \\times p > 3 \\times (1 - p) \\quad\\Longleftrightarrow\\quad p > \\frac{3}{3 + 30} \\approx 0.09 .$$\n",
    "\n",
    "If the probabilities are accurate, the cost-minimizing threshold is the false-alarm cost divided by the sum of the two costs, about 0.09 here, which is close to the 0.1 that the table shows to be cheapest. As with $k$ in Section {doc}`1501-knn-scikit-learn`, a careful analysis would choose the threshold on a validation set and confirm it on the test set."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "c032",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.401020Z",
     "iopub.status.busy": "2026-11-17T23:43:49.400946Z",
     "iopub.status.idle": "2026-11-17T23:43:49.402603Z",
     "shell.execute_reply": "2026-11-17T23:43:49.402275Z"
    },
    "tags": [
     "thebe-interactive"
    ]
   },
   "outputs": [],
   "source": [
    "### Exercise: A Threshold of 0.25\n",
    "#   1. Flag the test sessions whose p_buy is at least 0.25.\n",
    "#   2. Print the number flagged, the recall rounded to 3 decimals, and the\n",
    "#      total error cost with error_cost(y_test, ex_flagged).\n",
    "# Print exactly:\n",
    "# Flagged: 542\n",
    "# Recall: 0.607\n",
    "# Error cost: 7335\n",
    "### Your code starts here.\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "### Your code ends here."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "c033",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.403589Z",
     "iopub.status.busy": "2026-11-17T23:43:49.403523Z",
     "iopub.status.idle": "2026-11-17T23:43:49.407340Z",
     "shell.execute_reply": "2026-11-17T23:43:49.406965Z"
    },
    "tags": [
     "hide-input"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Flagged: 542\n",
      "Recall: 0.607\n",
      "Error cost: 7335\n"
     ]
    }
   ],
   "source": [
    "### Solution\n",
    "ex_flagged = (p_buy >= 0.25).astype(int)\n",
    "\n",
    "print('Flagged:', ex_flagged.sum())\n",
    "print('Recall:', round(recall_score(y_test, ex_flagged), 3))\n",
    "print('Error cost:', error_cost(y_test, ex_flagged))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c034",
   "metadata": {},
   "source": [
    "## Comparing with Nearest Neighbors\n",
    "\n",
    "```{index} comparing classifiers; logistic regression and nearest neighbors\n",
    "```\n",
    "\n",
    "```{index} KNeighborsClassifier; predict_proba\n",
    "```\n",
    "\n",
    "A nearest neighbor classifier (kNN, short for $k$-nearest neighbors) can also report a probability: `predict_proba` gives the share of the $k$ neighbors in each class. With $k = 11$, the possible probabilities are $0/11, 1/11, \\ldots, 11/11$. So you can compare the two classifiers on the same test set, both at the usual threshold of 0.5 and at a lower threshold suited to the costs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "c035",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.408457Z",
     "iopub.status.busy": "2026-11-17T23:43:49.408389Z",
     "iopub.status.idle": "2026-11-17T23:43:49.440683Z",
     "shell.execute_reply": "2026-11-17T23:43:49.440040Z"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>classifier</th>\n",
       "      <th>threshold</th>\n",
       "      <th>accuracy</th>\n",
       "      <th>precision</th>\n",
       "      <th>recall</th>\n",
       "      <th>error cost ($)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>kNN (k = 11)</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.888</td>\n",
       "      <td>0.667</td>\n",
       "      <td>0.554</td>\n",
       "      <td>8124</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>logistic regression</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.885</td>\n",
       "      <td>0.745</td>\n",
       "      <td>0.388</td>\n",
       "      <td>10728</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>kNN (k = 11)</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.808</td>\n",
       "      <td>0.436</td>\n",
       "      <td>0.832</td>\n",
       "      <td>4728</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>logistic regression</td>\n",
       "      <td>0.1</td>\n",
       "      <td>0.802</td>\n",
       "      <td>0.431</td>\n",
       "      <td>0.869</td>\n",
       "      <td>4221</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            classifier  threshold  accuracy  precision  recall  error cost ($)\n",
       "0         kNN (k = 11)        0.5     0.888      0.667   0.554            8124\n",
       "1  logistic regression        0.5     0.885      0.745   0.388           10728\n",
       "2         kNN (k = 11)        0.1     0.808      0.436   0.832            4728\n",
       "3  logistic regression        0.1     0.802      0.431   0.869            4221"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "knn = make_pipeline(StandardScaler(), KNeighborsClassifier(n_neighbors=11))\n",
    "knn.fit(X_train, y_train)\n",
    "p_buy_knn = knn.predict_proba(X_test)[:, 1]\n",
    "\n",
    "def summarize(name, probabilities, threshold):\n",
    "    flagged = (probabilities >= threshold).astype(int)\n",
    "    return {'classifier': name, 'threshold': threshold,\n",
    "            'accuracy': round(accuracy_score(y_test, flagged), 3),\n",
    "            'precision': round(precision_score(y_test, flagged), 3),\n",
    "            'recall': round(recall_score(y_test, flagged), 3),\n",
    "            'error cost ($)': error_cost(y_test, flagged)}\n",
    "\n",
    "pd.DataFrame([summarize('kNN (k = 11)', p_buy_knn, 0.5),\n",
    "              summarize('logistic regression', p_buy, 0.5),\n",
    "              summarize('kNN (k = 11)', p_buy_knn, 0.1),\n",
    "              summarize('logistic regression', p_buy, 0.1)])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c036",
   "metadata": {},
   "source": [
    "At the usual threshold of 0.5 the two classifiers are close in accuracy (0.888 for nearest neighbors against 0.885), but nearest neighbors catches many more buyers (recall 0.554 against 0.388) and so has the lower error cost (\\$8,124 against \\$10,728). At a threshold of 0.1, which the cost rule suggests, both become much cheaper, and logistic regression is the cheaper of the two (\\$4,221 against \\$4,728). Both then beat the \\$9,381 of flagging every session. So neither method wins everywhere: the exercise below shows nearest neighbors ahead at a threshold of 0.2. On this real data set the two methods are close, and the choice of threshold matters more than the choice of method. Logistic regression draws one smooth curve in each feature, which may not suit a feature such as page value, where most sessions sit at exactly 0 (Section {doc}`../14-classification/1401-classification`).\n",
    "\n",
    "Beyond the numbers, the two methods have different strengths:\n",
    "\n",
    "- **Logistic regression** gives smooth probabilities, coefficients whose signs explain the direction of each feature's effect, and fast predictions. It assumes that the probability changes smoothly and in one direction as each feature grows.\n",
    "- **Nearest neighbors** makes no such assumption and can follow irregular patterns, but it offers no coefficients to explain its predictions, needs scaled features, and must search the whole training set for every prediction.\n",
    "\n",
    "In practice, analysts often try both, compare them on validation data with the measure that matters to the business, and choose a threshold to match the costs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "c037",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.441972Z",
     "iopub.status.busy": "2026-11-17T23:43:49.441880Z",
     "iopub.status.idle": "2026-11-17T23:43:49.443515Z",
     "shell.execute_reply": "2026-11-17T23:43:49.443177Z"
    },
    "tags": [
     "thebe-interactive"
    ]
   },
   "outputs": [],
   "source": [
    "### Exercise: Compare at a Threshold of 0.2\n",
    "#   1. Flag test sessions at a threshold of 0.2, once with p_buy (logistic regression)\n",
    "#      and once with p_buy_knn (11-nearest neighbors).\n",
    "#   2. Print the error cost of each with error_cost(y_test, ex_flagged), and the name\n",
    "#      of the cheaper classifier.\n",
    "# Print exactly:\n",
    "# logistic regression cost: 6270\n",
    "# kNN cost: 5079\n",
    "# Cheaper: kNN\n",
    "### Your code starts here.\n",
    "\n",
    "\n",
    "\n",
    "\n",
    "### Your code ends here."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "c038",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-11-17T23:43:49.444459Z",
     "iopub.status.busy": "2026-11-17T23:43:49.444394Z",
     "iopub.status.idle": "2026-11-17T23:43:49.447906Z",
     "shell.execute_reply": "2026-11-17T23:43:49.447367Z"
    },
    "tags": [
     "hide-input"
    ]
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "logistic regression cost: 6270\n",
      "kNN cost: 5079\n",
      "Cheaper: kNN\n"
     ]
    }
   ],
   "source": [
    "### Solution\n",
    "ex_costs = {}\n",
    "for name, probabilities in [('logistic regression', p_buy), ('kNN', p_buy_knn)]:\n",
    "    ex_flagged = (probabilities >= 0.2).astype(int)\n",
    "    ex_costs[name] = error_cost(y_test, ex_flagged)\n",
    "    print(name, 'cost:', ex_costs[name])\n",
    "\n",
    "print('Cheaper:', min(ex_costs, key=ex_costs.get))"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "celltoolbar": "Tags",
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.13.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}
