BEST(Bayesian Estimation Supersedes the T-test)

baysean_ttest.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Bayesian estimation supersedes the t-test(BEST)\n",
    "\n",
    "This model replicates the example used in: Kruschke, John. (2012) Bayesian estimation supersedes the t-test. Journal of Experimental Psychology: General.\n",
    "\n",
    "https://docs.pymc.io/notebooks/BEST.html\n",
    "\n",
    "「Pythonで体験するベイズ推論」　Cameron Davidson-Pilon / 玉木 徹"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Running on PyMC3 v3.3\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/yasumizuyoshiaki/.pyenv/versions/anaconda3-4.2.0/lib/python3.5/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import pymc3 as pm\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "plt.style.use('seaborn-darkgrid')\n",
    "print('Running on PyMC3 v{}'.format(pm.__version__))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([<matplotlib.axes._subplots.AxesSubplot object at 0x118c6b9b0>,\n",
       "       <matplotlib.axes._subplots.AxesSubplot object at 0x118de2e80>],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "drug = (101,100,102,104,102,97,105,105,98,101,100,123,105,103,100,95,102,106,\n",
    "        109,102,82,102,100,102,102,101,102,102,103,103,97,97,103,101,97,104,\n",
    "        96,103,124,101,101,100,101,101,104,100,101)\n",
    "placebo = (99,101,100,101,102,100,97,101,104,101,102,102,100,105,88,101,100,\n",
    "           104,100,100,100,101,102,103,97,101,101,100,101,99,101,100,100,\n",
    "           101,100,99,101,100,102,99,100,99)\n",
    "\n",
    "y1 = np.array(drug)\n",
    "y2 = np.array(placebo)\n",
    "y = pd.DataFrame(dict(value=np.r_[y1, y2], group=np.r_[['drug']*len(drug), ['placebo']*len(placebo)]))\n",
    "\n",
    "y.hist('value', by='group')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "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>group</th>\n",
       "      <th>value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>drug</td>\n",
       "      <td>101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>drug</td>\n",
       "      <td>100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>drug</td>\n",
       "      <td>102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>drug</td>\n",
       "      <td>104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>drug</td>\n",
       "      <td>102</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  group  value\n",
       "0  drug    101\n",
       "1  drug    100\n",
       "2  drug    102\n",
       "3  drug    104\n",
       "4  drug    102"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "μ_m = y.value.mean()\n",
    "μ_s = y.value.std() * 2\n",
    "\n",
    "with pm.Model() as model:\n",
    "    group1_mean = pm.Normal('group1_mean', μ_m, sd=μ_s)\n",
    "    group2_mean = pm.Normal('group2_mean', μ_m, sd=μ_s)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "σ_low = 1\n",
    "σ_high = 10\n",
    "\n",
    "with model:\n",
    "    group1_std = pm.Uniform('group1_std', lower=σ_low, upper=σ_high)\n",
    "    group2_std = pm.Uniform('group2_std', lower=σ_low, upper=σ_high)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "with model:\n",
    "    ν = pm.Exponential('ν_minus_one', 1/29.) + 1\n",
    "\n",
    "pm.kdeplot(np.random.exponential(30, size=10000), shade=0.5);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "with model:\n",
    "    λ1 = group1_std**-2\n",
    "    λ2 = group2_std**-2\n",
    "\n",
    "    group1 = pm.StudentT('drug', nu=ν, mu=group1_mean, lam=λ1, observed=y1)\n",
    "    group2 = pm.StudentT('placebo', nu=ν, mu=group2_mean, lam=λ2, observed=y2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "with model:\n",
    "    diff_of_means = pm.Deterministic('difference of means', group1_mean - group2_mean)\n",
    "    diff_of_stds = pm.Deterministic('difference of stds', group1_std - group2_std)\n",
    "    effect_size = pm.Deterministic('effect size',\n",
    "                                   diff_of_means / np.sqrt((group1_std**2 + group2_std**2) / 2))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/latex": [
       "$$\n",
       "            \\begin{array}{rcl}\n",
       "            \\text{group1_mean} &\\sim & \\text{Normal}(\\mathit{mu}=101.17977528089888,~\\mathit{sd}=9.4895243570467)\\\\\\text{group2_mean} &\\sim & \\text{Normal}(\\mathit{mu}=101.17977528089888,~\\mathit{sd}=9.4895243570467)\\\\\\text{group1_std} &\\sim & \\text{Uniform}(\\mathit{lower}=f(),~\\mathit{upper}=f())\\\\\\text{group2_std} &\\sim & \\text{Uniform}(\\mathit{lower}=f(),~\\mathit{upper}=f())\\\\\\text{ν_minus_one} &\\sim & \\text{Exponential}(\\mathit{lam}=0.034482758620689655)\\\\\\text{difference of means} &\\sim & \\text{Deterministic}(\\text{group1_mean},~\\text{group2_mean})\\\\\\text{difference of stds} &\\sim & \\text{Deterministic}(\\text{Constant},~\\text{Constant},~\\text{group1_std_interval__},~\\text{Constant},~\\text{Constant},~\\text{Constant},~\\text{group2_std_interval__},~\\text{Constant})\\\\\\text{effect size} &\\sim & \\text{Deterministic}(\\text{group1_mean},~\\text{group2_mean},~\\text{Constant},~\\text{Constant},~\\text{group1_std_interval__},~\\text{Constant},~\\text{Constant},~\\text{Constant},~\\text{Constant},~\\text{group2_std_interval__},~\\text{Constant},~\\text{Constant},~\\text{Constant})\\\\\\text{drug} &\\sim & \\text{StudentT}(\\mathit{nu}=f(\\text{ν_minus_one}),~\\mathit{mu}=\\text{group1_mean},~\\mathit{lam}=f(f(\\text{group1_std})))\\\\\\text{placebo} &\\sim & \\text{StudentT}(\\mathit{nu}=f(\\text{ν_minus_one}),~\\mathit{mu}=\\text{group2_mean},~\\mathit{lam}=f(f(\\text{group2_std})))\n",
       "            \\end{array}\n",
       "            $$"
      ],
      "text/plain": [
       "<pymc3.model.Model at 0x118e93550>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 解釈\n",
    "\n",
    "t分布は３つのパラメータ$\\mu$, $\\sigma$, $\\nu$　がある。\n",
    "\n",
    "- $\\mu$ : 平均値。正規分布からサンプルする。\n",
    "- $\\sigma$ : 分散。一様分布からサンプルする。\n",
    "- $\\nu$ : 外れ値の観測しやすさ。指数分布を+1shiftした分布からサンプルする。\n",
    "\n",
    "このt分布をgroup1, group2に対して推論する。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Auto-assigning NUTS sampler...\n",
      "Initializing NUTS using jitter+adapt_diag...\n",
      "/Users/yasumizuyoshiaki/.pyenv/versions/anaconda3-4.2.0/lib/python3.5/site-packages/pymc3/model.py:384: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  if not np.issubdtype(var.dtype, float):\n",
      "Multiprocess sampling (4 chains in 4 jobs)\n",
      "NUTS: [ν_minus_one_log__, group2_std_interval__, group1_std_interval__, group2_mean, group1_mean]\n",
      "100%|██████████| 2500/2500 [00:06<00:00, 391.42it/s]\n"
     ]
    }
   ],
   "source": [
    "with model:\n",
    "    trace = pm.sample(2000, cores=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x540 with 6 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pm.plot_posterior(trace, varnames=['group1_mean','group2_mean', 'group1_std', 'group2_std', 'ν_minus_one'],\n",
    "                  color='#87ceeb');"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x360 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pm.plot_posterior(trace, varnames=['difference of means','difference of stds', 'effect size'],\n",
    "                  ref_val=0,\n",
    "                  color='#87ceeb');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "difference of meansは95%以上が0を超えているので、優位にgroup1のほうが高い事がわかる。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZ8AAAEdCAYAAADTmmpHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAF9pJREFUeJzt3XuUXWWZ5/FvkQRImWAMBmgFEq4PdHQ1El2gBgmoqChjoKeBRrQBofHSOEizQBGksV3tOGoGtEdRIEZbMYByUQcVhkESbmK8QHN7BggJF7kmRBMSAUnNH3sXOSkqdUnqvKdO1fezVlbO2WdfnvOey2+/++x6d0dXVxeSJJW0WasLkCSNPoaPJKk4w0eSVJzhI0kqzvCRJBVn+EiSihvb6gIkSetERBdwJ/Ai0AV0An8CPpqZi3qZ/1+AV2fmPw1yO9cAR2Xm05tc9EYwfDRgEbEF8G1gZ6oPw8cz876IOBT4MvBwPevZwG+AHwPjgRMz846ImAm8NTO/uIH1/yNwNLAWGAd8JjN/uRF1fgTYDjgf+GxmfiwilgB7ZOafG+Y7pp72qQGscx4wPzN/voHHdwT+JjN/Mth6+9nuLOAjmXnkUK5Xw94BjaEQEacCXwPePITbeOcQrmvQDB8NxgnAqszcNyIC+HfgXcAM4LTM/FH3jBFxGFX43AB8OCJOBv4b8MHeVhwRR1J9GN6emS9ExE7Agoh4w8bumWXm48DHNmbZjXAgsAcwpOEjRcRYYEdgeR+z7RER1wN/BTwBHJmZj0XE+4AzgM2BbYDvZOZZEfHternrI+LgzHy499U2j+Gjwfhr4GcAmZkRsWc9fQbwhjpgbgNOB1ZR9XrGA88CRwFXNPY8ejgROCUzX6jX/2BE7JWZy+pex9b1v/cCpwH7AWOAOZl5Wd2rOg94BvgLcGtETKPqrexbb+Ob9bQngH9o3HhEnFTX2FUv89Xeiqx7IqcDz1P1AOcD/x34FNAZETcDDwJfBTqAZcBxwBuAL9bL/R9gVmYeUK/zp8BZwC7Ax6l6fV3AoT22/W1g17pNz8vM/9hAW6r9XR8Ra4EpwJ+BnwLH9jH/zsA+mflURFwJHB8Rnwf+GfiH+gjFa4CHIuK8zDy27vkfsLE7d5vKEw40GL8H3hcRHRGxL/DaiBgDXAucBLwNmAB8hOoLdluqUPkW1Rfp7RHxzYg4rZd1vwZY3DghM5c13P2/mfkWYF9gp8ycCRwAfCYiJgHfAP4+M99B9eXfm29k5v7AEqpeHAAR8dfAEcBMqlCbXffsNmQq8Ld1Ladl5otUAXRxZv4YuIDqkOQs4GqqsATYMjP3y8xzgC0jYmpE/BXV8frfAbsD762f291UvcruGidSte9hwLupfg/QyHVAZv4N1c5WJ3BzZj7Zx/zXZuZT9e3bgW0ysws4BJgREWcDc6h2iF7RxLoHzJ6PBmMusCewELgJ+E1mvhgRczNzBUBEXAX8bWauBU6up32aqldyJlVI/UtE7J6Z/69h3UuBHYA/dk+IiHcBd9R3s/7/9VQfpl/W98cB04BtG9Z3E1UPodHzmXlrfftmqkN8v67vv44qUK6r778K2K1hmz39Z2b+BfhLRKzp5fE9ga/X+TUOuK/HcwC4CPgQ8BzV72gATwLfiYhVVIfwbumeOTNX1j3LbwFbAd/bQG0aQTLzdxHxSeDCiLg1M5dExNVUO2sAn63/f6FhsS6gIyJeAfwOuILqMzsXmE0VQC1nz0eD8SbgunrP/DJgcUR0AHdExPb1PG+nOtkAgIjYBojMXEi1B9d9Bk/Pva+5wFn18W0iYnfgQtbt4a+t/78XuL7uVRwIXAo8ADzacBjwTb3UvnlE7FXf3o/qbKJuCdxFtbc5C5jHutDrTW+j8a5l3ecpgQ/V6zqN6pBJ43OA6nDd+6h6hBdHxCuBc4AjgeOBNTR8SdQ9pBmZeSjV3vD/6G4rjWyZ+QOqHZFz6/sHZ+Ze9b8f97HoblQ7KmfWJ8LsD2xBdbgaqs/WuOZV3jfDR4NxH3ByRNwC/CvVbzRdVF+Wl0fEDVQBc0HDMmcCn69vfx34BdWPorc3rjgz5wO3AjdGxAKq3sDRvRxq+AmwKiIWUoVcV2aupDq8992IuI6qF9PTc8BJ9bq3oQq27m3fTtXruTEiFlF9aB8deLMA8J/A++sTJz5a13Ij1eG4lwVZZq6q2+Ceuv4/UfXYbqHaS13Dur1bgMeB7erflK4Fvlz3vjQ6/BPwnvpowEDdQbXjc29E/Bb4L1SHc7uPClxO9Z5/3ZBWOkAdXlJBklSaPR9JUnGGjySpOMNHklSc4SNJKs7wkSQV598JDNJTT61c7/TACRO2YNWq51pVzkax5jKsuXmmTJnYsj+U7Pkd0K1d2q4ZNvTc+3qd7PlsorFjx/Q/0zBjzWVY8+gymttuY5674SNJKs7wkSQVZ/hIkoozfCRJxRk+kqTiDB9JUnGGjySpOMNHklSc4SNJKs7wkSQVZ/hIkoozfCRJxRk+kqTiDB9JUnGGjySpOMNHklSc4SNJKs7wkSQVZ/hIkoozfCRJxRk+kqTiDB9JUnGGjySpOMNHklSc4SNJKs7wkSQVZ/hIkoozfCRJxRk+kqTiDB9JUnGGjySpOMNHklTc2FYXII1kj6xYwylX3sXS5auZOrmTObOns/2k8a0uS2o5w0ej3smX38lNDy5v+nYeXLaaQy/69ZCu8607Tebcw143pOuUSjB81DRHzFvE4mWrW13GiHbTg8t501cWtLqMIbPz1p1ccswbW12GCjB81DRD+SUyaVInK1a0V5BNmtTJQecuYMny1XR1QUcHTJvcyaXD+Mu1HdtZ7ckTDqQmmjN7OtMmd7JZHTxzZk9vdUnSsGDPR2qi7SeNH9Y9HalV7PlIkoozfCRJxRk+kqTiDB9JUnGGjySpOMNHklScp1pLTeb4btLLGT5SrcQYb80Y3w0c403tx/BREY7z1lztNMab47cJDB8VsqlfNu045lh3zYfPW9Q247u1YzurPXnCgdRkju8mvZw9H6nJHN9Nejl7PpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFTcsT7WOiDHAJcCFmfnzVtcjSRpawy58ImIX4LvA9sCFLS5H2iQOKjrydb/GDy1fw46Tx/saD9CAwycidgXmAS8AS4FpmTkrIpYC9wJ3A+cBc+v1dgGfyMzbI+LxzNyuXs984HxgGjAbmAi8GvhcZv4ImAAcD5w+gJq665kKbAHMBw4BdgTen5kPRMQXgP2AMcCczLwsIvYHzqY67DgBOAp4HvgB8DCwC3BbZn50oO2jkaGZg4s2Y1BRBxRtvVOuvOul4ZOWLF/NKVfe5R8VD8Bgej5fAv4tM6+OiBOowgNgB2DvzFwWET8EzsvMqyJiL+AioK9X4RXAO4EpwG0RcVVm3g4QEQOta0lmnhAR5wM7ZebBEXEOcEhEZD1tZkRsCdwaEdcC04GjM/MPEXEG8HfA94HdgYOA1cDiiNguMx9v3NiECVswduyYl+6PGbMZkyZ1DrTWYaHZNR/8tRu578lVTVu/1mmnAUV72m2bCVx90sxWl7HJltbBA9DVVd1X/wYTPnsCN9e3FwIfqG8/nZnLGuZZAJCZv4+IHXpZT0fD7Rsycy3wREQ8QxVCjw2iJoDf1v+voOqBATwDbAm8HpgREb+sp4+jCs1Hga9GxCrgtcBN9eP3Z+ZKgIh4rF7Helatem69++04EGOza774g3sP+TrbtZ0POndB2wwqCuXbeWO3NWXKxCGuZONNndy53ms8dXJ77Yy2ymDOdrsTeHN9e9+G6Wsbbt9DdYiLuufT3WsYFxETImJzql5Htxn1vNsCWwFPDqKebl19PHYvcH1mzgIOBC4FHgAuAI7NzGOAP7AuEPtalzRoDio68nW/xmM6OnyNB2EwPZ/TgbkRcSrwR6rfWno6Fbignmcc8OF6+rnArcBiqt+Lum0XEdcBrwQ+lpkvDrL+/vwEmBURC6l+27kiM1dGxPeAhRHxLPAE8Joh3q4EOKjoaND9Grdj77yVOrq6BrazHxEfAH6VmfdHxPHAWzLzuI3dcEQcA+yRmZ/a2HW0wlNPrVyvwdrxDWfNZVhz80yZMrGj/7mao+d3QLd2abtm2NBz7+t1GkzP52FgfkSsBl5kXa+mqepDddf08lBm5oklapAkDa0Bh09mLqDvM9cGJTPnDXC+54FZQ7VdSVLrObyOJKk4w0eSVJzhI0kqzvCRJBVn+EiSijN8JEnFGT6SpOIMH0lScYaPJKm4YXclU2kk8oqm0voMH6nWzKuYNvKKppLhoyY5Yt4iFi8bnSP8tkI7XNF05607ucTLS6hm+KgphvpLph2Hq2+s+fB5i9riiqbt2M5qT55wIBXgFU2l9dnzkQrwiqbS+uz5SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScWNbXUBao1HVqzhlCvvYuny1Uyd3Mmc2dPZftL4VpclaZQwfEaIky+/k5seXL5Ryz64bDWHXvTrQS3z1p0mc+5hr9uo7UmS4VPYEfMWsXjZ6laXscluenA5b/rKgiLb2nnrTn5x8tuKbEtSGYZPYZcc88ZWl8CkSZ0cdO4ClixfTVcXdHTAtMmdXDoMapM0OnjCwSg1Z/Z0pk3uZLM6eObMnt7qkiSNIvZ8RqntJ423pyOpZez5SJKKM3wkScUZPpKk4gwfSVJxho8kqTjDR5JUnKdaj0IPLV/N8d9d5LhuklrG8BkBHNdNUrsxfJpkpIzh1psS47rtvHXnsBiKSFJzGD5NMpy/OP/+u7/hgaefdVw3SS3jCQej0DePnuG4bpJayp7PKLSjPR1JLWbPR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUNu7PdIuLtwOeBF4AngQ9l5sj8a01JGqWGY8/n68DszHwbcB9wfIvrGXEeWr6aw+ctYp85Czh83iIeWbGm1SVJbeuRFWs4fN4i9vjsL/w8DcKAez4RsSswj6pHshSYlpmzImIpcC9wN3AeMLdebxfwicy8PSIez8zt6vXMB84HpgGzgYnAq4HPZeaPgFmZ+URDfX/uo6bueqYCWwDzgUOAHYH3Z+YDEfEFYD9gDDAnMy+LiP2Bs6nCdwJwFPA88APgYWAX4LbM/OhA26eVHNtNap1TrryLJctX09UFS5av5pQr7/Lv6AZgMIfdvgT8W2ZeHREnUIUHwA7A3pm5LCJ+CJyXmVdFxF7ARUBfr8IrgHcCU4DbIuKqzHwMICIOAw4AzuqnriWZeUJEnA/slJkHR8Q5wCERkfW0mRGxJXBrRFwLTAeOzsw/RMQZwN8B3wd2Bw4CVgOLI2K7zHy8cWMTJmzB2LFjXro/ZsxmTJrU2WeBB3/tRu57clU/T6N9lBjbrTe7bTOBq0+aWXy7G2sg743hph1rbrWldfAAdHVV99W/wYTPnsDN9e2FwAfq209n5rKGeRYAZObvI2KHXtbT0XD7hsxcCzwREc9QhdBjEfFJ4L8C787MDfZ8ar+t/19B1QMDeAbYEng9MCMifllPH0cVmo8CX42IVcBrgZvqx+/PzJUAEfFYvY71rFr13Hr3J03qZMWKvt9sF39w736eQlntOLZbdzv319bDyUDeG8NNu9Q8ZcrEVpfwkqmTO1/q+XR0VPfVv8H85nMn8Ob69r4N09c23L6H6hAXdc+nu9cwLiImRMTmVL2ObjPqebcFtgKejIjP1Ot4R2Y+PYC6uvp47F7g+sycBRwIXAo8AFwAHJuZxwB/YF0g9rWuEcOx3aShM2f2dKZN7mRMR4efp0EYTM/ndGBuRJwK/JHqt5aeTgUuqOcZB3y4nn4ucCuwmOr3om7bRcR1wCuBj1H99nM2VW/mZxEBcElmfmMQdTb6CTArIhZS/bZzRWaujIjvAQsj4lngCeA1G7n+tuTYbtLQ2X7SeC495o1t02scLjq6uga2sx8RHwB+lZn3R8TxwFsy87iN3XBEHAPskZmf2th1tMJTT61cr8Ha8Q1nzWVYc/NMmTKxo/+5mqPnd0C3dmm7ZtjQc+/rdRpMz+dhYH5ErAZeZF2vpqnqQ3XX9PJQZuaJJWqQJA2tAYdPZi6g7zPXBiUz5w1wvueBWUO1XUlS6w3HPzKVJI1who8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4gwfSVJxw+4y2irnkRVrOOXKu1i6fDVT69F4t580vtVlSRoFDJ8RYlOuZgqDv6KpVzOVtCkMnwKOmLeIxctG1mi3zb6a6c5bd3KJl32QRizDp4Dh9iXaPfz54fMWrXcFxna4oqmkkcETDkax7iswekVTSaXZ8xnFuq/AKEml2fORJBVn+EiSijN8JEnFGT6SpOIMH0lScYaPJKk4w0eSVJzhI0kqzvCRJBVn+EiSijN8JEnFGT6SpOIMH0lScYaPJKk4w0eSVJzhI0kqzvCRJBVn+EiSijN8JEnFGT6SpOIMH0lScYaPJKk4w0eSVJzhI0kqzvCRJBVn+EiSijN8JEnFGT6SpOIMH0lScYaPJKk4w0eSVJzhI0kqzvCRJBVn+EiSijN8JEnFGT6SpOIMH0lScYaPJKk4w0eSVFxHV1dXq2uQJI0y9nwkScUZPpKk4gwfSVJxY1tdwHAUEfsAX8zMWRGxKzAP6ALuBD6emWsj4mzgvcBfgJMz87Ye6zgE+Gz9+NzMvKANav4kcDzwVD3pxMzMVtZcz7crcEVmvr6XdewLnEf1nK7JzHOaVe8Q1nwo8GXg4XrS2Zl5QytrjogvATOpvhO+1fP9Wrqd201jG7e6llIiYhwwF5gGbAF8PjN/PNDl7fn0EBGnARcCW9aT5gBnZuZ+QAfw/ojYG9gf2Ac4EvhfPdYxDvifwEH1fP8YEdsO55prM4APZeas+l8zg6ffmuv5PgjMB6ZsYFXnA0dRfXHuExFvaIOaZwCnNbRzM4NnIO+NA4BdM/PNVO14ekS8qseqirVzu+mljUeLo4Fl9Xvp3cC/D2Zhw+flHgAOa7g/A+j+cvgZ8A6qD+A1mdmVmQ8BYyOi8YtmT+D+zHwmM58HbgTeNsxr7l7u0xFxY0R8uon1DrRmgGeoQvNlImIrYIvMfCAzu4BfNCzXDJtcc8Nyx0XEwoj4SkQ08wjEQGq+BTiuntYFjAFe6F6gBe3cbnq28WhxGXBWfbuDqlc8YIZPD5n5Ixo+eEBH/YEDWAm8EtgK+GPDPN3Tu/X3+JAaopqh2lv/CHAgMDMi3tecigdcM5n508x8dgOr2Qr4U8P94dDO/dUMcC1wEtUOyQSqNm+KgdScmX/OzGfqHvt3qA67rWpYpmg7t5te2nhUyMxVmbkyIiYCPwTOHMzyhk//1jbcngisoPogTuxlerf+Hm+2QdccER3AuZn5dN1b+99AyUMrvdXcn+HYzgMxNzMX1yFwFcOgnevDbD8H7s7ML/RYptXtrGEqInYArgf+IzMvHsyyhk//fhcRs+rb7wEWAjcB74qIzSJiR2CzzHy6YZl7gN0iYnJEbE61h3vLMK95K+DOiJhQB9GBwG9aXHOfMvNPwPMRsUtd87sGstwQGnTNdZ13RMT29aS30+J2jojxwHVUofivPRcYBu2sYaj+Hfsa4PTMnDvY5T3brX//DFxQh8g9wA8z88WIWEgVKJsBHweIiKOACZn5rYg4herY+GZUH+pH26DmM6j2Yp4DrsvMq1tZ84ZmjIgDgZmZ+TmqQ1bfp/qd4prM/FWJYmsbVXNEHA9cHhFrgLuBpp4J2UNvNX8C2Bk4ISJOqOc7FtiJ4dHOGp7OAF4FnBUR3b/9vCcz1wxkYYfXkSQV52E3SVJxho8kqTjDR5JUnOEjSSrO8JEkFWf4SJKKM3wkScUZPpKk4v4/tXFZcWJoHvEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pm.forestplot(trace, varnames=['group1_mean',\n",
    "                               'group2_mean']);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "pm.forestplot(trace, varnames=['group1_std',\n",
    "                               'group2_std',\n",
    "                               'ν_minus_one']);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "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>mean</th>\n",
       "      <th>sd</th>\n",
       "      <th>mc_error</th>\n",
       "      <th>hpd_2.5</th>\n",
       "      <th>hpd_97.5</th>\n",
       "      <th>n_eff</th>\n",
       "      <th>Rhat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>difference of means</th>\n",
       "      <td>1.002127</td>\n",
       "      <td>0.44729</td>\n",
       "      <td>0.004676</td>\n",
       "      <td>0.107068</td>\n",
       "      <td>1.862148</td>\n",
       "      <td>6899.0</td>\n",
       "      <td>1.000230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>difference of stds</th>\n",
       "      <td>0.932048</td>\n",
       "      <td>0.44503</td>\n",
       "      <td>0.005439</td>\n",
       "      <td>0.091474</td>\n",
       "      <td>1.812321</td>\n",
       "      <td>6332.0</td>\n",
       "      <td>0.999873</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>effect size</th>\n",
       "      <td>0.596257</td>\n",
       "      <td>0.28138</td>\n",
       "      <td>0.002947</td>\n",
       "      <td>0.056113</td>\n",
       "      <td>1.137876</td>\n",
       "      <td>6419.0</td>\n",
       "      <td>1.000327</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                         mean       sd  mc_error   hpd_2.5  hpd_97.5   n_eff  \\\n",
       "difference of means  1.002127  0.44729  0.004676  0.107068  1.862148  6899.0   \n",
       "difference of stds   0.932048  0.44503  0.005439  0.091474  1.812321  6332.0   \n",
       "effect size          0.596257  0.28138  0.002947  0.056113  1.137876  6419.0   \n",
       "\n",
       "                         Rhat  \n",
       "difference of means  1.000230  \n",
       "difference of stds   0.999873  \n",
       "effect size          1.000327  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pm.summary(trace,varnames=['difference of means', 'difference of stds', 'effect size'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "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>mean</th>\n",
       "      <th>sd</th>\n",
       "      <th>mc_error</th>\n",
       "      <th>hpd_2.5</th>\n",
       "      <th>hpd_97.5</th>\n",
       "      <th>n_eff</th>\n",
       "      <th>Rhat</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>group1_mean</th>\n",
       "      <td>101.540341</td>\n",
       "      <td>0.384866</td>\n",
       "      <td>0.003951</td>\n",
       "      <td>100.753486</td>\n",
       "      <td>102.268671</td>\n",
       "      <td>6679.0</td>\n",
       "      <td>1.000184</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>group2_mean</th>\n",
       "      <td>100.538214</td>\n",
       "      <td>0.225966</td>\n",
       "      <td>0.002765</td>\n",
       "      <td>100.128883</td>\n",
       "      <td>101.011508</td>\n",
       "      <td>6716.0</td>\n",
       "      <td>1.000200</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>group1_std</th>\n",
       "      <td>2.123269</td>\n",
       "      <td>0.435399</td>\n",
       "      <td>0.005634</td>\n",
       "      <td>1.312711</td>\n",
       "      <td>2.992706</td>\n",
       "      <td>5659.0</td>\n",
       "      <td>0.999986</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>group2_std</th>\n",
       "      <td>1.191221</td>\n",
       "      <td>0.159981</td>\n",
       "      <td>0.002027</td>\n",
       "      <td>1.000017</td>\n",
       "      <td>1.506498</td>\n",
       "      <td>6866.0</td>\n",
       "      <td>0.999932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ν_minus_one</th>\n",
       "      <td>0.977157</td>\n",
       "      <td>0.513533</td>\n",
       "      <td>0.006583</td>\n",
       "      <td>0.080532</td>\n",
       "      <td>1.999856</td>\n",
       "      <td>6040.0</td>\n",
       "      <td>0.999839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>difference of means</th>\n",
       "      <td>1.002127</td>\n",
       "      <td>0.447290</td>\n",
       "      <td>0.004676</td>\n",
       "      <td>0.107068</td>\n",
       "      <td>1.862148</td>\n",
       "      <td>6899.0</td>\n",
       "      <td>1.000230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>difference of stds</th>\n",
       "      <td>0.932048</td>\n",
       "      <td>0.445030</td>\n",
       "      <td>0.005439</td>\n",
       "      <td>0.091474</td>\n",
       "      <td>1.812321</td>\n",
       "      <td>6332.0</td>\n",
       "      <td>0.999873</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>effect size</th>\n",
       "      <td>0.596257</td>\n",
       "      <td>0.281380</td>\n",
       "      <td>0.002947</td>\n",
       "      <td>0.056113</td>\n",
       "      <td>1.137876</td>\n",
       "      <td>6419.0</td>\n",
       "      <td>1.000327</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                           mean        sd  mc_error     hpd_2.5    hpd_97.5  \\\n",
       "group1_mean          101.540341  0.384866  0.003951  100.753486  102.268671   \n",
       "group2_mean          100.538214  0.225966  0.002765  100.128883  101.011508   \n",
       "group1_std             2.123269  0.435399  0.005634    1.312711    2.992706   \n",
       "group2_std             1.191221  0.159981  0.002027    1.000017    1.506498   \n",
       "ν_minus_one            0.977157  0.513533  0.006583    0.080532    1.999856   \n",
       "difference of means    1.002127  0.447290  0.004676    0.107068    1.862148   \n",
       "difference of stds     0.932048  0.445030  0.005439    0.091474    1.812321   \n",
       "effect size            0.596257  0.281380  0.002947    0.056113    1.137876   \n",
       "\n",
       "                      n_eff      Rhat  \n",
       "group1_mean          6679.0  1.000184  \n",
       "group2_mean          6716.0  1.000200  \n",
       "group1_std           5659.0  0.999986  \n",
       "group2_std           6866.0  0.999932  \n",
       "ν_minus_one          6040.0  0.999839  \n",
       "difference of means  6899.0  1.000230  \n",
       "difference of stds   6332.0  0.999873  \n",
       "effect size          6419.0  1.000327  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pm.summary(trace)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 事後分布の生成\n",
    "\n",
    "サンプルしたパラメーターを用いてt分布の事後分布を作成する。sample_ppcを使う。\n",
    "\n",
    "https://docs.pymc.io/notebooks/posterior_predictive.html"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 50000/50000 [00:30<00:00, 1618.75it/s]\n"
     ]
    }
   ],
   "source": [
    "ppc = pm.sample_ppc(trace, samples=50000, model=model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x11cfb9e48>"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# sns.kdeplot(ppc['drug'], bw = 10)\n",
    "# sns.kdeplot(ppc['placebo'], bw = 10)\n",
    "plt.hist(ppc['drug'], bins=1000, density=True, label='drug', alpha=.5)\n",
    "plt.hist(ppc['placebo'], bins=1000, density=True, label='placebo', alpha=.5)\n",
    "plt.xlim(70,150)\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 普通のT検定（分散が違うことがわかっているのてWelch）をやってみる。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([101, 100, 102, 104, 102,  97, 105, 105,  98, 101, 100, 123, 105,\n",
       "       103, 100,  95, 102, 106, 109, 102,  82, 102, 100, 102, 102, 101,\n",
       "       102, 102, 103, 103,  97,  97, 103, 101,  97, 104,  96, 103, 124,\n",
       "       101, 101, 100, 101, 101, 104, 100, 101])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Ttest_indResult(statistic=1.622190457290228, pvalue=0.10975381983712831)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from scipy import stats\n",
    "\n",
    "res = stats.ttest_ind(y1, y2, equal_var = False)\n",
    "res"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "優位性示せず。\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/yasumizuyoshiaki/.pyenv/versions/anaconda3-4.2.0/lib/python3.5/site-packages/seaborn/categorical.py:598: FutureWarning: remove_na is deprecated and is a private function. Do not use.\n",
      "  kde_data = remove_na(group_data)\n",
      "/Users/yasumizuyoshiaki/.pyenv/versions/anaconda3-4.2.0/lib/python3.5/site-packages/seaborn/categorical.py:826: FutureWarning: remove_na is deprecated and is a private function. Do not use.\n",
      "  violin_data = remove_na(group_data)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x11bde8550>"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAQ4AAAHaCAYAAADxHqecAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3XmcVNWd9/HPrb33jabZV/GAgIpEUXEhqOCCkrjFZ5KYjEaTPMZsk+1JzGjWSV5PlmcyycTE6JhtdKJGIyjuiqgoKoiKcFhkX7tpeq+97vNHVTfNInR1V91Tt+v3fr18WbeWWz+o4lvnnHvuuZZt2wghRDY8pgsQQriPBIcQImsSHEKIrElwCCGyJsEhhMiaz3QBx9PY2C6HfYQwpL6+wjra/dLiEEJkTYJDCJE1CQ4hRNYkOIQQWZPgEEJkTYJDCJE1CQ4hRNYkOIQQWZPgEEJkTYJDCJE1CQ4hRNYkOIQQWZPgEEJkLW9nxyqlZgE/1VrPUUqdBPwesIANwGe01gml1E3AZ4EE8EOt9eJ81SOEyJ28tDiUUt8A/gCEMnf9GPi21np2ZvtypdQw4IvAbGA+8G9KqWA+6hFC5Fa+uiqbgCt7bV+ltX5RKRUAhgGtwBnAy1rrqNa6FdgInJyneoQQOZSXrorW+iGl1Lhe20ml1FjgGdKhsRq4JHO7WztQdfi+ysuD+HzefJQphOgnx1YA01pvBSYppT4D/AJ4CKjo9ZQKoOXw13V0RJ0pUAhxhPr6iqPe78hRFaXUo0qpSZnNdiAFrADOVUqFlFJVwBTgXSfqKSTxeBy5KJZwG6cOx/4EuFcp9TxwPemB0j3Ar4BlwHPAd7TWEYfqKQidnZ18/ONX8ec/32O6FCGyYhX6r91gXqx4z57d3HrrzQA88MAiw9UIcSRZrLgApVIp0yUI0S8SHAYlk0nTJQjRLxIcBqVSEhzCnSQ4DEokJDiEO0lwGJRMJkyXIES/SHAYJGMcwq0kOAxKJKTFIdxJgsMgCQ7hVhIcBklwCLeS4DBIgkO4lQSHQYlE3HQJQvSLBIdB0uIQbiXBYVA8Li0O4U4SHAbF4zHTJQjRLxIcBkmLQ7iVBIdBvYOj0NdFEaI3CQ6DYrGDXRUZKBVuIsFhUO8xDum2CDeR4DCod4sjFpPV3IV7SHAYdGhwyBEW4R4SHAb1bmVIcAg3keAwSLoqwq0kOAySropwKwkOg6LRKLaV/ggkOISbSHAYFI3FsL1BQKafC3eR4DAoFoth+4I9t4VwCwkOg2KxKPgCmdsSHMI9JDgMisXiPV0VCQ7hJhIcBiXiMexMi0PGOISbSHAYFE/Esb3dwSHnqgj3kOAwxLZtEvGDwSFdFeEmEhyG9JxG7/GB5ZEWh3AVCQ5DuoPC9niwPF4JDuEqEhyG9FwawfKCxyML+QhXkeAwJB7PBIXlAUtaHMJdJDgMSSbTwWFbHrA8PdtCuIEEhyHJZDJ9w/JgW56D20K4gASHIT1jGpkWh4xxCDeR4DAklerd4rCkxSFcRYLDkFQqlb5hWYCFbaeM1iNENiQ4DEml0hdgsi0LLKtnWwg3kOAw5GALw8KWFodwGQkOQw695KN1sOsihAtIcBQIy7JMlyBEn0lwFAQZ3xDuIsFhzOEtDGlxCPeQ4BBCZE2CwxCPJ/NXb9tY2Ae3hXAB+bYa0h0UFimwJTiEu8i31RCv15u+kWlx9GwL4QISHIb07qpgp6TFIVxFvq2GHGxxpLsq0uIQbiLBYYjP50vfsFNYdgqfz2+2ICGyIMFhSHdwWHYK7CQ+n7Q4hHtIcBhysMWRBDuF1+szW5AQWZDgMKSna5JKYSeT+P3SVRHuIcFhSHdQWHYSUhIcwl0kOAzxer3pM2KTccAmEAiYLkmIPpPgMMSyLHx+P1YyfT0VOaoi3ESCwyC/P4CVjAJIi0O4igSHQX6/HyuRvkp9ICAtDuEeEhwGBQK9WxxBw9UI0XcSHAb5/QGsRDRzW1ocwj0kOAwKBoOQ7O6qSItDuIcEh0HB4MEWhwyOCjeR4DAo4A/0rDTq90twCPeQ4DCod/dEWhzCTfJ2ZpVSahbwU631HKXUqcB/AEkgClyvtd6rlLoJ+CyQAH6otV6cr3oKUe+wkOAQbpKXFodS6hvAH4BQ5q5/B27VWs8B/g58Uyk1DPgiMBuYD/ybUqqoRgglOIRb5avFsQm4EvhzZvs6rfXuXu8ZAc4AXtZaR4GoUmojcDLweu8dlZcHB+1aFRUVZT236+urqa4uNViNEH2Xl+DQWj+klBrXa3s3gFLqbOALwHmkWxmtvV7WDlQdvq+Ojmg+SiwIqdTBizCFwwmgy1wxQhxFfX3FUe93bHBUKfUx4E7gMq11I9AG9K6qAmhxqp5CcGhXpah6acLlHFl2Sin1CdKDoHO01s2Zu1cAP1JKhYAgMAV414l6CkXv4JDFioWb5D04lFJe4FfANuDvSimApVrr25VSvwKWkW75fEdrHcl3PYVE5m4It8pbcGittwBnZjZrP+A5dwF35auGQuf3yzqjwp1kAphBcmKbcCsJDoNk1S/hVhIcBkmLQ7iVBIdBEhzCrSQ4DOq5KJMQLiPBYZBcvU24lQSHQdLiEG4lwWGQBIdwKwkOg2SauXArCQ6DJDiEW0lwGCTBIdxKgsMgj0eCQ7iTBIdBHo/89Qt3km+uQRIcwq3km2uQBIdwK/nmGiTBIdxKvrkGSXAIt5JvrkGWZR3/SUIUIAkOgyxL/vqFO8k31yCPR1ocwp0kOIyS4BDuJMFhkAxxCLeS4DBIBkeFW0lwGCXBIdxJgsMgaXEIt5LgEEJkTYJDCJE1CQ6DbNs2XYIQ/SLBYZQEh3AnCQ4hRNYkOAySropwKwkOgyQ3hFtJcBglySHcSYLDoFRKgkO4kwSHQbadMl2CEP0iwWGQDI4Kt5LgMCiVkhaHcCcJDoMkOIRbSXAYJMEh3EqCw6BUKmm6BCH6RYLDoGRSWhzCnSQ4DEompcUh3EmCwyDpqgi3kuAwKJGQ4BDuJMFhUCIRN12CEP0iwWGQjHEIt5LgMCiRSJguQYh+keAwSLoqwq0kOAyKxyU4hDtJcBgUi8VMlyBEv0hwGCQtDuFWEhwGxePS4hDuJMFhUDQaNV2CEP0iwWFQJBLpuS1zOoSbSHAYFIsdbHFI60O4iQSHQZFI7+CIHOOZQhQWCQ6DIpFwz+1wOHyMZwpRWCQ4DAqHu3rdluAQ7iHBYVDXIcHRdYxnClFYJDgM6uzsIuUvBSQ4hLtIcBjU0dFBKljRc1sIt5DgMKirqwM7ExydnRIcwj0kOAxJJBJEIxFSwXJAWhzCXSQ4DOluYdi+EJY/REdHu+GKhOg7CQ5D2tpagXRw2L5Qz7YQbiDBYUhLSwsAtr+EpC/Usy2EG+QtOJRSs5RSLxx23y+VUp/rtX2TUuoNpdSrSqkF+aqlEPW0OPwhUr4QByQ4hIvkJTiUUt8A/gCEMtv1SqklwBW9njMM+CIwG5gP/JtSKpiPegpRS8sBAFL+Emx/Ca2ZbSHcIF8tjk3Alb22y4E7gD/3uu8M4GWtdVRr3QpsBE7OUz0FZ//+JvB4wRvEDpQSDncdcpq9EIXMl4+daq0fUkqN67W9GdislLqk19Mqgd4jgu1A1eH7Ki8P4vN581GmUW1tLRAsB8siFUgfko3FOhg2rNZwZUIcX16Co4/agIpe2xXAER39jo7BuU7Frl27SfjS083tzLTzzZu3U1k5xGRZQhyivr7iqPebDI4VwI+UUiEgCEwB3jVYj6P2NTZiB9Kti+5JYE1NjSZLEqLPjB2O1VrvAX4FLAOeA76jtS6KTn40GqG1pZlUqBIAO1AGloc9e3YbrkyIvslbi0NrvQU487D77jhs+y7grnzVUKj27t0LQCqYDg4sD4QqJDiEa8gEMAN2794F0NPiAEgEKtm5a6epkoTIigSHAbt27QB6tThIh8iePbtIpVKmyhKizyQ4DNi2bWv6UKwv0HNfsqSGeCzG3r17DFYmRN9IcBiwectmEqGaQ+5LlaaPsGzbtsVARUJkR4LDYfF4nN27dpIsPSw4QtWAxZYtm80UJkQWJDgctm3bVlKpJKmSw2aIen1QUsXmze+bKUyILEhwOGzDBg1AsnzoEY/FS4eg16/Dtm2nyxIiKxIcDlu/fh0EStOTvg6TLB9KR3ubDJCKgifB4bC169YSL60HyzrisWR5PZAJFyEKmASHgw4caKapce9RuykAqZIaLF+AtWvXOFyZENmR4HDQ22+/BUCycsTRn2B5iJc38NbqtxysSojsSXA4aPXqVVj+kp45G0eTqBxJU+NeGecQBU2CwyG2bbP67beIVQw76vhGt+7WyNtvr3KqNCGyJsHhkC1b3qettYVE5chjPi8VqoJgOStXvuFQZUJkT4LDIa+9thywSFaPPvYTLYtY1RjeWr2KcDjsSG2if/7yl3v52McWsmbNO6ZLcZwEh0OWv/oyyYph2P6S4z43UTOWRDzOW2+tdKAy0V+rV68klUqxefMm06U4ToLDATt3bmfXzh3Ea8b26fnJigbwh3jttVfyXJkYiHgsvR5ua2vxXYVPgsMBy5e/DECiekzfXmB5iFWN4fU3VhCNFsVqiq504EAzkLnURZGR4Mgz27Z5/oVnSVYMx84sStwXibqJxKIRVqx4NY/Vif5qb2+nK5wO9X37iu/QuQRHnq1fv459e/cQG3JCVq9LVgyDYAUvvPBsnioTA7FnT3r5xwp/it2ZFd2KiQRHni1d+hyW10eiZlx2L7QsorUTeOed1ezfvz8vtYn+27ZtKwCzhsZoa++gtbW4rv0rwZFHkUiEZS+9SKx6LHj9Wb8+PuQEbNtm6VJpdRSazZvfJ+SDmfXRnu1iIsGRRy+/vJRIuIt4verX6+1QFcnK4Tz51BKSyWSOqxMDsV6/x4SKOBMqk1gcXGelWEhw5Ilt2yxZ8hh2aS3J8oZ+7ydWP4Xm/U2sWvVmDqsTA9HZ2cGWrVs4sSpOic9mdEWKNWveNl2WoyQ48mTjxvVs3bqZaL36wHNTfE0b8DVtOOZ+EtVjIFDKk08+no8yRT+sWfMOtm1zUk0cgKnVUfS6tUQixXPoXIIjTx5/fBGWN0C87oOPpvibNuA/TnDg8RAdciJvrV7JLrlgU0FYufINSnxwQlUCgJPrYiSSSd55p3iWQ5DgyIP9+5t45ZWXiA6Z1K9B0cPFh04By8Pjjz+ag+rEQCSTSd54fTkn10bxZf71qOoEpX6Kas6NBEceLFmymJSdItZwUk72Z/tLiNdO4Lnnn6G9vT0n+xT9895779La1s7pQ2M99/k8MKMuyusrXiEejxuszjkSHDkWDod56uklxKvHYgcrcrbfWMNU4rEYzzzzRM72KbK3bNkLhHxwSl3skPvPbIjS2RVm1ariWA5BgiPHXnjhGcJdXcQapuZ0v6nSWpKVI3jssUVF86tWaCKRCK8uX8bpQyIEvYc+Nq0mTmWQopnpK8GRQ8lkkn88+jCp8qGkKvp/CPaDRIdNp7X1AC+9tDTn+xbH9+qrLxOORDl3ePSIx7wemD00zMo3X6el5YCB6pwlwZFDr776MvubGokOm56X/ScrR2CX1vLwIw/JVe0NePqpxxlWZqOqE0d9/PwREZKpFM8//4zDlTlPgiNHbNvm4YcfxC6p7vvp89myLCLDprN71w5ZWtBhW7ZsZv2G9cwd3vWBS8aOKEsxpSbO0089Puhn+kpw5Mi7776dnvDVMPWYixEPVKJmPATLeeQfD+XtPcSRlixZRMDLUbspvV0wMkJjU9OgD3YJjhz5x6N/h0AJ8bqJ+X0jj4fI0Knode+xYcP6/L6XAKC1tYVlLz7P7IYIZf5jX9d35pAYtSFYvOhhh6ozQ4IjB7Zt28rqt1YSrZ8CHl/e3y9efyKWL8Cjj/497+8l0rOAE4kEF485/uLRXg/MH9XJe2vXDOpgl+DIgUWLHgavj9jQyc68oddPZIjitddekQs35VlHRwdLHn+UmfUxhpf2bUB6zogIZX548MH78lydORIcA9TScoBly5YSq5sEvpBj7xtvmIqNxZIlixx7z2K0aNHDhCMRPjKuq8+vKfHBJaO7WLnyjUF7uv1xg0Mp5VVK3aiU+r5Sao5SaogThbnFs88+RTKZSJ9P4iA7UEq8ZizPPvdMUZ2V6aSmpkYWLfo7ZzZEGVOR3VGSi0aFqQrCH++9C9s+9riIG/WlxfE7YCxwEVAB/CmvFblIMpnkiScfJ1k5glRJtePvHx96EpFwF8uWveD4ew92tm1z9913QirBNRP63troVuKDq8d3oNfrQTmvoy/BMVFr/a9AWGu9CKjKc02u8frrr9JyoJno0NyczJatZPlQ7LI6Hl+yeFD+qpn00ktLeeONFVw1vpP6kv5Ntjt3eBRVneCP995FY+O+HFdoVl+Cw9fdPVFKVQAyZTHjqaefgGA5yepRZgqwLKL1k9mxfSvr1w/OvrQJO3Zs5/e/+zWTqpLMH9X/bqDHgs9MbicVD/PLX/xkUJ1j1JfguA14GfgQ8Crw/bxW5BLNzft5953VRGsngmVujDleOx48Pl588TljNQwmLS0H+OlPvoffjnLLtDa8A/xoG0pTfGZyOxs2buC3v/3VoDlV4Lh/LVrrpVprBUwEpmmtn85/WYXvpZeWYts28Syvl5Jz3gDx6jG89PKyQfWLZkJ7ezs//MF3aW7ay5ent1IbzM0/8jOGxrh6Qnos6r/+6/eDolt53NlKSqnnAbvXNlrruXmtygWef+E5UuX12CHzQz7xuol0bXiflSvfYNass0yX40qNjfv40Q+/y749u/jS9DYmVR39RLb+unxsmM64xZInHqOrq4vPf/6L+Hz5nyyYL32p/HOZ/1vATODU/JXjDjt37mDH9q3ExpzZ/53YNlasCysZw79vXfoSCv08xyVZNRL8JbzyyjIJjn7Qei0//9mPiXa28PVTWplck9vQgPRHe90JXZT6bB568XkONO/nS1/+OlVVzh+Ny4XjBofWuveo2zql1I15rMcV3norfamCRPXofu/D37gOb7QNgNDWVwC7/3NBLA/xypG8tXoVyWQSr9d7/NcIUqkUjz32D/76l3upCyX52oxWRpfn76xWy4KF48PUhlLcu/Ztvv61W/nKV7/FlCm5XfTJCX3pqtzca3M40PcrJw9Sq1a9iV1SPaClAX0t24/YHsgkskTVSLr2b+T99zcyaVL/LgBVTHbv3sWdv/0V761dw8z6GDdN6aDU58zYw7nDo4wtT/Afa2xuv/1bXHrpFVx33ScJhZybeTxQfemqDO91OwJcm6daXCEajbLmvXeJ1544sB2lEsfezlKiaiQAb721UoLjGOLxOI899g8e+Nt/4yPODZM7OH94NJ8rIRzVmIok3/9QM3/bVMZjjz3KGyuWc+NN/5sZMz7kbCH99IHBoZTq/pdx+Jk6gfyVU/g2blxPIh4nUTnCdCmH8oVIlQ3hnXfe5ppr/pfpagqObdu8/vqr/OmPf2Dvvn3MHBLjetVBTdDcEY4SH3xKdTJraJR71tv8+MffY8app/GpT3+GkSP73w12wrFaHL/7gPttoGiPqmzZshmAVFmd4UqOlCitY8vWzdi2jeX0T2gBW7fuPe777z/x3to1jCxP8fVTOpheVziHrifXJPjx6c08vSPEP9as5F+++gU+PHceV199HXV1hfc9g2MEh9b6w0e7XylV1C2OrVs3YwVKsP2lpks5Qqq0lnCjZv/+JoYMqTddjnGbN2/ivvv+zKpVb1IVhOtP7ODDI6IDntSVDz4PXDImwuxhUR7ZXMrzzz3B0qXPcPHFC1i48Gqqqswf9u+tL4OjnwW+CvhJH5KNAwPs4LvX5i2biYdqTJdxVKmSdF1bt24p6uDYuHE9Dz54P2+++Tplfrh2YicXjTrykgaFqDJgc73q5JIxYR7eXMLixY/w1JOPMW/+ZVx++ZXU1BTGd68vg6O3AHNITz1/APhyPgsqdLt27SRVNWHgO0rGCIVCLFiwgMWLF9ORjB3/NcfbZSY4du3awcyZpw94f26zdu0a/v7Q//DW6lWU++Gq8V1cNDqSt6MlL+1ON77PGT7wz+5w9SUpbj6pk8vHhnl0aymLFz/CE08s5sILL+aKK640/sPQl+DYrbXerZSq0Fq/oJS6Pe9VFah4PE4sGsH2D/ywmZWIseCKBXzhC1/Atm3+tujJgRfoDYBlFdVlIm3bZvXqVfz9oftZu24tFQG4ZkInF46KUJLniZkv7k5/D/IRHN2Gl6X47EkdfGRcF4u2lvDUE4t56qnHOe+8uXz0o9cwfLiZQfq+/NW2KKU+AtiZbkvRLuTT2dkBgO0LDnhfti/A4sXp0+Efe+wxbF8OxkwsC8sX6qlzMLNtmzfeeI2HHryfTe9vojYEn5jUyfkj3NElyVZDaYrPTOnkI+PDPL4txNIXn+GFF57l7LPP5aqrrmX06LGO1tOX4KgHpgHfAr4G3JrXigpYR0cmOLw5GB/2Boh0NfPQQ5nLHFTkZuqx7Qv01DkY2bbNihXLeeBv/83WbVsZWmpzw+ROzhl28Orxg9mQUIrrT+xi4bgwS7aV8OxrL/LKKy9yxhlncc01/8TYseMcqaMvwfF14J+BJcAjwPt5raiARaPptRlsj99wJR/M9vh66hxMursk9/33H3l/8/sMK7O5eUoHZzXECvIoSb5VBWyuO6GLy8aEeXJ7iKffXM6KFcuZPft8PvaxjzNs2PDj72QA+nKuypvAm0qpGuC3wAZg4G11FwoG031aa4CzPPPJSiV76hwstm3byj333MmaNe8ypMTmpimdzB4WxSNTVagI2Fw9MczFYyI8trWEp5cvZfkry7ho3qVcd93HKSvLzxkifTkcey7waeB00kdVvpaXSlyg+1yCwg6OuKvOeTiWrq4u/va3v7JkySJKvTafPLGTOSMi+IuwhXE85X6bj53QxfzRYR7ZUsqTTy7mlZdf4BOfvJHzz5+Lx5Pbv7S+dFW+DNwFfEZr7f4VSAYgFCpJ30gWzqzDIyQTB+t0sQ0b1vPLX/wbTU1NzBkR4eqJXVQc5ypqTrJtOBD1EE5YPLczyIdHOH++y9FUB20+rdIB+6f1Sf7zP/+dV199mS984StUVFTm7H360lW5Kmfv5nIlJSVYHg9WokDHEFJJ7ESUsrIy05UMyFNPLeGee+6kOpDitpmtOV9UJxee2xlkbzh9+OZeXY5twwWjjn1dWSeNq0hy22ktPLMjxP1vvcHXv3Yr3/jmd5kwITcr1kmjLwter5f6+gY8kVbTpRyVJ5Je32PEiJGGK+m/Z599krvu+k+mVUf4wYeaCzI0AFbtDxxzuxB4LJg3OsJ3T2uB8H5++IPb2LFj+/Ff2Jd952QvRWT0qNH4MgvwFBpPpAWAkSMNrbo+QFqv5Xe/+w3Ta+N8aXo75QXUNTlcLGkdc7uQjK9M8q1TWvDEO/nRD79LNDrwlpEER5ZGjhyFFWkFu/BWq+5uCQ0f7s4WxyOPPEi53+bWaW0FPycjnLAIhUJcffXVhEIhwonCDQ5ITyD77JQ2mvbv56WXlg54fwX+8RSeMWPGQiqJJ1x43RVP136GDBnqyqMqyWSSN998nbMbwoRcsIZvV8JiwYL0KQOXXXYZXQUeHABTa+I0lKYn0A1U3j4ipdQs4Kda6zlKqROAe0mv5fEucIvWOpU57+UyIAF8WWu9Il/15Mrkyemrtnnbd5MqLYwzFQGwbQIde5l69tmmK+kXj8eDz+t1zS9Zqc8+5JSBBoeWHRwIy0r/FwwOfBpWXj4npdQ3gD8A3T99vwBu01qfS/rU/IVKqdOA84FZwHXAb/JRS641NAyjtq4eb/se06UcwhM+gB2PMG3ayaZL6RfLspgwYSLL95UQLszx0EOU+GwikQgPPfQQkUiEEhcEx9oDPvZ0Wjk5spKvgN8EXNlreybQ3bFaAlwInAM8pbW2tdbbSF9q0hWLSEyfNp1Ax570wfwC4W3fDcDUqdMNV9J/n/r0TbRE4b/WlZMovCEkV9sf8XC3rqR+yBAuueTyAe8vL10VrfVDSqlxve6yek0eayd94epKYH+v53Tf39h7X+XlQXy+wjrdcdasM1i69Dk8nU2kyvuZdR7fsbez5GvdScOw4UyaNG5A+zHp9NNncOONn+Huu/9ANGXx+ZPaC3a8I+C1j7ldSHZ2evnZ29WECfHj2/6VhoaBd7Gd+lh6/35UAC1AW+b24fcfoqOjcCbVdJsy5RQ8Hi++A1uI9TM4EtWj8bXuOGS73xJRfG27OOO8hbS0dPV/PwXg4osXYll+7r77t3zn9VpunNzGSXm4QNJAzaiL8XavuRsz6vK3Jkd/pWx4cnuIBzeXUVpWwR23/YCRI8dn9R2prz/6JUCcGotapZSak7l9CbCM9IWs5yulPEqpMYBHa93kUD0DUl5ezvTpJxNs2drv7kq8fjLJYCUpX4jI2LOJ10/udz2+lu1gpzjrrNn93kchmT//Ur73vZ/grRzGT1ZVcffaMlqihXXUYu7IKA0lSSr9KT6tOpg7srB+4Da2+vjBymru21jGKTPO4P/+7NeMHz8xZ/t3qsXxL8BdmYWO1wIPaq2TSqllwHLSAXaLQ7XkxJlnzmb16lV4ws2kSvuxErVlYQdKsSklPrT/oQHgO7CF6po6Jk6cNKD9FJIpU6bys5//mvvv/ytLHn+U5ftCXDy6i8vGFMZApGVBTTBFTZCCCo3dnR4eeL+MNxoDVFVWcuutn+Hcc+fkfNV7q9CvnN3Y2F6QBba2tnLTzdcTGTqV2Oj+re9Zsu5xAMKTL+1/IYkoFavv59KLL+Of//mm/u+ngO3evYv77/8zr7zyEmV+uGhkF/NGR4zPLP3xyvRJY98+zfxM4h0dXhZtLeHVfUGCgSBXLLyKBQs+QknJwE54rK+vOGriFOjQU+GrqqritBkf4s133iU2aiZYZmYg+Jvfh1SSOXMG76Vuhg8fwVe+8k2uuOIqHnzwPh55YwVP7Chl7ogw80ZHqA0W7yGYTa0+Fm8t4c2mAMFggAULLmPhwivzfjFrCY4B+PCHL+TNN1fgbd1JciCDmwMQaNrI6DHjctp/LVQTJ55n+li4AAAX5ElEQVTAN7/5XbZt28rDDz/Akpdf5MntJZzZEOWS0WHGVOTvgtGFJGXDyqYAS7aVsKHVR1lpKddcs5BLLlmQ01Pnj0WCYwBOO+1DlJWVE2/aYCQ4POEDeDobueCahY6/t0ljxozlS1/6Gtdd9wkee+xRnnvuSV7eE2RqbZyLR4eZXhsflKuDRRKwbE+IJ3eUsq/LYmh9PTfc8FHmzLlwwF2SbElwDIDf7+f88+fy+JLFRONhbL+zH56/cT0er5dzzjnf0fctFA0Nw7jhhpu59tp/4umnl7Dk8Uf5+eoWRpSlmD+qi9nDogTyOAXovOHOrMvSHPXw9PYQz+8uoSsOk06YxCeuuJIzzjgLr9fMHCcZHB2g7du38dWv3kJk1OnEh2c3a3NAg6OpBBWr/4ezTj+dr371m9m/fhCKx+MsX/4Sixf9nc1btlAZhHkju5g70vxAan/s6PCyZFuIV/aFSNkWs2adxYIFH0WpgR2Fy4YMjubJ6NFjUJNPQm/RxIdNw6n143zNWyARZd68Sxx5Pzfw+/2cd96HOffcOaxZ8w6P/uMhHnxrJYu2lTJneJhLx4SNXp2+rza0+li0tYS3mgIE/H7mzZ/PZZctpKFhmOnSekhw5MD8eZegf/VzvG27SFY5sxZGsHEdDcNGuPrclHyxLItp005m2rST2bp1C48++neefukFnttZwvnDIywYG6Y2VHhHYnSLj39sKeXdZj8V5WVce+1C5s+/lMrKwrrgNEhw5MSsWWdTVvY74o3akeDwdDXj6djH/CtvyPnEnsFm7Nhx3HrrV7n22n/i4Ycf4Pnnn+GF3SEuGBnminHhglgAeVu7l/s3lfFus5+qygo++cmrueiiSxwf8MyGBEcOBAIB5s69iEWLHyEa68IO5OByjsfg37cOr8/HnDkX5PV9BpOGhmF87nO3ctVVH+PBB+/nqeefYdmeEi4f08m80WYuudAc9fDgphJe3hOitLSU66+/jnnzLnHFdXHcsm5KwbvooovBtvE3rc/vGyXjBJo3Mfvscx07Zj+Y1NcP5fOf/yI/+9l/oKbN5H82lXHb67XoFud+Q1M2PLU9xLdeq+G1pnIuv+JKfv2bu7n88o+6IjRAWhw5M3z4CKZPP5V31mtiw0/O20xS//5NkIwzb94ApqkLxowZy7e/fQerVr3JXb//NT9aaTF3ZIT/dUJnXi9avafLw+/WVrKp1cspJ5/KTTffUlCDnn0lLY4cmj//Uoh24m3Zcfwn94dtE2zSjB4zlhNPVPl5jyIzY8ZMfvHL33L55R/l+V0h7nizhp2d+UmO5XsC/OsbNexNVPClL32N79z2fVeGBkhw5NTMmadTWVVNoFHnZf+erv1YnfuZP+8SGRTNoVAoxPXX38B3vvN9OjxV3P5GNauacndh8ZQN920o5bfvVTBu4mR+9vPfcM4557v6M5TgyCGfz8cFcy/C17YDK9qR8/37G9fh9wc455w5Od+3gFNOmcH//dlvGD12Av/vnUqW7hr4or6JFNy5ppwl20u4+OIF3PG9n1BXNyQH1ZolwZFjF1wwLzNIuiG3O07GCTZvZvbsc11/icdCVlNTw+13/ISTT57B3evKeXl3/6/QlrLh92vLeXVfkI9//FPccMPNxqaI55oER441NAxj+vRTCe7fkNPFjP3Nm7GTcS688OKc7VMcXUlJCd/4xm1MnTqNu9ZV8G5z/7ot928s5dW96dD4yEeudnXX5HASHHlwwQUXQbSjZ+XxXAjs38Dw4SNlUNQhgUCAb3zju4wcOZrfvldJcyS7fyor9gV4ItM9Wbhw8F23XYIjDz70oVmESkpz1l2xIq142vcyd+6Fg+pXq9CVlpbyL1/7P8SsAL9fV97nBmRL1OIeXcGkEyZx/fWDc3avBEceBINBzj3nPPwHtkJy4Ktf+5s2YlkW55334RxUJ7IxcuRoPvnJG3mv2c8re/s23vGXDWXEbR+3fvFr+P25OzpTSCQ48mTOnAsglcB3YOvAdmTbBJvf5+STT6W2th+LIosBu+iii5k4YSJ/e7+C2HEWGdvQ6mPFviAfvfJahg8f4UyBBkhw5MmkSYoh9UPx739/QPvxdOyDaDvnnjsnN4WJrHk8Hj7xyRs4EIHndh17SviD75dRVVnJ5Zd/1KHqzJDgyBPLsjj3nPPxte/Ciof7vR9/8/v4/H5OP/3MHFYnsjVt2smcNGUqT2wv+8DLU77f5mXtAR8LP3INoZA7zjnpLwmOPJo9+zywbXzNm/u3AztFoGULM087ndLS/J5xK47vioVX0RyBNxqPPtbx5PYSSkIhLrxwnsOVOU+CI4/Gjh3HiJGj8B/Y0q/Xe9v3QiycDiBh3IwZM6kfMoTnj9JdaY9bvN4YZM6HL6SkZPCHvARHnp115my87Xv71V3xHdiKz+9nxoyZeahMZMvj8XDBhRez9oCfpvCh/3Re2xsgkcrMHC4CEhx5dsYZZwI23pbtRzwWHzKJ+JAPuGyjbRNo3capp8wY9P1lN+leUf7VfYd2V5bvCzFm9GjGjh1voizHSXDk2fjxE6mtG4K/5cjDsokhk0h8QHB4uvZDtIMzzjgr3yWKLDQ0DGPihIm82XQwzFuiFhtbfJx1dvF0KSU48syyLGadcSa+tt2QSvT5db6WbWBZzJzZv+vSivw5/Yyz2NTqpTWWnhG6en8AG4rqyJcEhwNmzPgQpBJ42/f0+TX+tp2cMPHEglzhuth1jzl1n/z2TrOfmupqxowZa7IsR0lwOOCkk6bh8/vxte7s2wsSETwdjZx2mgyKFqJx4yZQVlrKugN+bBvWtQaZfvKMQXlOygeR4HBAMBhk6knT8Lf1LTh8rbsAOPXU0/JZlugnj8eDmjyVDW1B9oU9tEVh8uSTTJflKAkOh5x66mlY4ZY+rQzmbdtJSWkZEyac4EBloj9OOGESuzst1rX4e7aLiQSHQ6ZOPRmgT+McgY69TJs6fdCsFjUYjRs3ARt4ZU8Qr8fDqFFjTJfkKAkOh4wdO46S0rLjLu5jRTsg0sa0aXJpx0I2atRoANa2+BnW0DBoT5//IBIcDvF4PEybOp1Ax95jPq87WOSasIVt6NCGnsHQhuGjDFfjPAkOB02dOg0ibVixzg98jrd9LyWlZYweXTyH9tzI6/VSW1MNpK8OV2wkOBx04omTAfB2NH7gc/xdjSg1GY9HPppCV16RnmNTU1NjuBLnybfTQePGTcDr9eHt/IDgSMawug5w4iRZkNgNrMxlPotxkp4Eh4P8fj/jx0/E27nvqI97O5sAZCVzl+ie71VeXm62EAMkOBx24okKb9d+sI9cRqq7JTJx4olOlyUGoBjW3zicBIfDxo+fAMkEnkjbEY95OpupGzK0KH/B3CwYHPilIt1GgsNh48ZNAMDT1XzEY/5IMxMnTHC6JDFAgUD/LxPpVhIcDhs5chRery+93kZvyTiEWxk/fqKZwsQAFM/Jbd0kOBzm9/sZOWo03vChLQ5P+ACQnmEq3KI7MHJ3jWC3kOAwYMzoMfiih45xeMItAEV3zsPgIC0O4YBRo0ZDpD3dPcnwRlrx+XwMHdpgsDLRP9LiEA4YOTJ9boMn0tpznyfcwrDhI+SMWFfpDgxpcQgHHC04fLE2RmfOuBSi0ElwGNDdHfFE29N32CmItNPQMNxgVUL0nQSHAcFgiMqq6p7gsGJdYKdoaBhmuDLRH0W01GgPCQ5Dhg0b3hMc3f+X4HAnu/jGRiU4TGkY2oAvnl6Xw8oERzGu6+BuMo9DOKyubgh2tBNsG0+8C4Da2jrDVYlsFNPlEA4nwWFIXd0QsFNYiQhWrJPyisqiPOfBzexi7KNkSHAYUleXbl1YsU48sS7qpLXhYsXX8pDgMKSmphZIH1HxJrp6gkS4ibQ4hMOqqtIL3XoSETyJCFVVxbdupXAvCQ5DuteptOJh7FiYqqriW7dysCjGMVIJDkOCwSDBUAhPtA3slASHixXjGKkEh0Hl5ZU9p9NXVFQarkb0l7Q4hKMqyst7Jn/JOqPuJS0O4aiKigo8iQgAZWUSHMI9JDgMKi0t67ldVlZ2jGeKQlSMLY1uEhwG9Q6L0lJpcbhNMY5tdJPgMKikpOSot4W7pFJJ0yU4ToLDoFDoYFgU40V9BotYLGa6BMdJcBjUu5Uha426VzQaNV2C43xOvZFSKgj8FzABaANuAeqAfwcSwFNa6+85VU8hCAZDpksQA9A9OBqJhM0WYoCTLY6bgA6t9ZnArcCvgTuBfwLOAWYppWY4WI9xchr94NDV1WW6BMc5GRwnAUsAtNYaOB0Iaq03aa1t4EngQgfrMU6Cw+3STY7Ozg7DdTjPsa4K8BawQCn1CDALqAI29Xq8nXQ35hDl5UF8vsHZ/6+pqei5XV1darAS0R/JRHpQNBYLF93n52Rw3ANMAZYBLwOrgd6zniqAlsNf1NExeAeeYrGDM4haWoqvuet2LQfSX9fGxv2D9vOrr6846v1OdlVOB57VWp8DPACsB2JKqYlKKQuYTzpUiobP52Rui1yKx+O0d6YXm25pOWC4Guc5+c3dAPxAKfUd0i2LG4ExwF8BL+mjKq85WI9xEhzu1dp6sHHcvL/JYCVmOPbN1Vo3ceTg5y7gTKdqKDQSHO61f/9+AGqDSZqb9xuuxnkyAcwgmfTlXt1hMbEyQXtHR9HNHpXgMEiCw72am9Pdk4lVicx2cbU6JDgM8njkr9+t9u9vwu+FMeXpE9wkOIRjPB5pcbhVc3MzNUGoCaYy2xIcwiHS4nCvAweaqfYnqA6kMtvFdUhWvrkGFfO1R92urfUAlYEkpT4brwVtba2mS3KUBIdBEhzu1d7eRrnfxrKgPADt7e2mS3KUBIdBEhzuFQ5HKPGlTxko8dmEw4NzyvkHkeAwSILDnWzbJhaPE/CkgyPgsYtuMR8JDiGylEqlB0S9mdz3WTbJZMJgRc6T4BBigFKAZRXXP6Xi+tMKkQMejwfLskhkVkVI2VbRzQKW4BAiS5ZlEfD7iSbTfZVo0lN0q9RLcBhkF/OlwFyutLSESCIdHF1J65Cr8hUDCQ6DJDjcq6Kikva4h5QNnTG76C4aLsEhRD9UVdfSFvfQGbdI2VBdXW26JEdJcBgkLQ73qq6uoSXm40As/U+opqbWcEXOkuAwqHs+gHCf2to6DkThQESCQzjMtiU43KqubgjJFGzt8PVsFxMJDoNSKemquFVdXR0AG1t9WJa0OISDpKviXt0tjE1tPqoqK4tu4WkJDoMkONyrtjbd4miPe6ipqTNcjfMkOAxKpZKmSxD9VFlZ1XN2c22RjW+ABIdR0uJwL6/XS2Vm0ld1dY3hapwnwWGQBIe7VVZWHfL/YiLBYVAyKV0VN/P6/QBUVBz9wsyDmQSHQdLicLv0GEexneAGEhxGSYvD3brnchRji6O4Dj4XGDmq4m433vg5zjxzNqeeOtN0KY6T4DAokZDgcLP6+qHMmXOB6TKMkK6KQdLiEG4lwWGQtDiEW0lwGCQtDuFWEhwGJRLFdS0OMXhIcBgkLQ7hVhIcBskYh3ArCQ6DpMUh3EqCwyBpcQi3kuAwSFocwq0kOAySc1WEW0lwGCTBIdxKgsOg3sEhp9gLN5HgMKh3WEhwCDeR4DCod4tDui3CTSQ4DDq0qyLBIdxDgsMg6aoIt5LgMKj31erlcpDCTSQ4DOrdPZEWh3ATCQ6DDm1xSHAI95DgMOjQ7ol0VYR7SHAYdTAserc+hCh0EhwG9W5xSHAIN5HgMEqCQ7iTBIcQImsSHEKIrElwCCGyJsFRIGSMQ7iJBIdRVs8tj0c+CuEe8m01yLIO2TJVhhBZk+AwyOqVHJbkhnARCQ6DDg0O+SiEe8i31aDe4xoyxiHcRL6tBvVuZXg80lcR7iHBYZC0OIRbybfVoN5hIWMcwk3k22qQ1+s96m0hCp3PqTdSSvmBPwLjgCRwE5AA7iV9mui7wC1a66JZCku6KsKtnPy2Xgr4tNZnA98HfgT8ArhNa30u6RlQCx2sxzhpcQi3cjI41gM+pZQHqATiwExgaebxJcCFDtZjXO+wkBaHcBPHuipAB+luyjpgCLAAOE9r3X12VztQdfiLysuD+HyD89e4tDTUc7u6utRgJUJkx8ng+ArwpNb6/yilRgPPAYFej1cALYe/qKMj6lB5zkskDg7ntLR0GaxEiKOrr6846v1Oto8PAK2Z282AH1illJqTue8SYJmD9Rjn8QzOlpQY/JxscfwSuEcptYx0S+PbwBvAXUqpALAWeNDBeozzemVcQ7iTY8Ghte4Arj3KQ+c7VUOh8XqdzG0hckd+8gwarIO+YvCT4DBIxjiEW0lwGOTzSVdFuJMEh0EyW1S4lQSHQRIcwq0kOAySoyrCrSQ4DJJ5HMKt5JtrkBxVEW4lwWGQjHEIt5LgMEiCQ7iVBIdB0lURbiXBYZAMjgq3km+uQdJVEW4lwWGQLBco3Eq+uQZJcAi3km+uQRIcwq3km2uQHFURbiXBYZC0OIRbyTfXIAkO4VbyzTVIgkO4lXxzDbIsy3QJQvSLBIdB3S2OioqjX/RGiEIlK8kYVFFRyawzz2bO+ReYLkWIrFi2bR//WQY1NrYXdoFCDGL19RVH7U9LV0UIkTUJDiFE1iQ4hBBZk+AQQmRNgkMIkTUJDiFE1iQ4hBBZk+AQQmRNgkMIkTUJDiFE1iQ4hBBZk+AQQmRNgkMIkTUJDiFE1iQ4hBBZk+AQQmRNgkMIkbWCXwFMCFF4pMUhhMiaBIcQImsSHEKIrElwCCGyJsHhAKVUSCm1xXQd4viUUluUUqEc7OcOpdTnclFTIZLgEEJkTQ7H5olSqhz4K1ADbATmAluAfUAtcB8wSWv9rcwv3Dqt9Til1BnAb4D2zHMjWutPO/8nGPyUUp8GPgJUAEOA7wM/ByYDJwC/ALyZxz6vtX5FKXUj8PnM/Y9qrW9XSl0DfBVIAi9lPtM7gLMAP1AKfFFrvUIp9XHgy0AU2ADcrLWOO/RHzhlpceTP54B3tdbnAb/rdf99WusLSX/JjuZO4NNa67nApjzXKKAMuAiYRzooui+LOhX4F631BcBPgX9WSg0FvgWcC5wGBJVSY4DvARdorc8BRiqlLsrsY3Pmc7wRuFMpVZd57tzMc1uAzzrxh8w1CY78ORFYAaC1fg3o/lXRR3lu78vsjdBar8ncXpa/8kTGUq11Smu9FzgA1Gfu3wl8Vyn1R+Bq0i2HCaR/DMJaa1tr/S1gWOY1jyulXgBOAiZm9vEiQObzHJZ5/RqtdXuvx6fm+w+YDxIc+fMe6aYqSqkZpL94AKnM/yPA8Mzt03q9brtS6qTM7TPzXaRgJoBSqgGoJN09BPgVcLvW+lPAO6TDfRMwWSkVzLzmQWAvsB24SGs9B/gP4NXMPs7IPG86sA3YDJyklCrLPH4+sD6ff7h8keDInzuBCUqpl4BbSPdpe3sCGJd5/FqgLXP//wbuUUo9Q/qL57r+r8sMU0o9CzxG+u++uwv5F+ABpdQy0q3HEVrrRtLdlqVKqeXASq31VtJdnKVKqdeASzgYBuOVUs+R/i58VmvdBNwOPK+UepX02MlvHflT5pgMjhYYpdQtwN+01o1KqR8CMa31903XNRhlBkcnZ7ocIgu+4z9FOGwv8JRSqgNoBT5luB4hjiAtDiFE1mSMQwiRNQkOIUTWJDiEEFmT4BBCZE2CQwiRNTkcK7KilCoB/gSMID1j8jzSE566T967DLib9PRqL/ALrfX/ZKZjf05rvS5zuvkw4F7gAWA3MApYorX+jqN/INEv0uIQ2bqZ9Mlbs4E7gIbM/d0n790ENGqtzwYuBH6olBpyjP2NAz4NnA7MVUqddoznigIhwSGyNQV4BUBrvQ5ozNyvez3efXJXO+lzdiYeto/eJ/Wt1lo3a62TwGuAylPdIockOES23uXgyXsTSZ9vAQdP3ltL+rRzlFIVwHTSJ3d90El9U5RSpUopLzCLdNCIAifBIbJ1N+mT814k3VWJHPb474G6zMl7LwDf01rvI3226X8qpZ4kPfbRLUZ6nOM14B9a69X5LV/kgkw5F1lRSp0NlGutn1JKTQKe0Fof3hXp677GAfdrrWX5AJeRoyoiW+8D9ymlbie9xsgthusRBkiLQwiRNRnjEEJkTYJDCJE1CQ4hRNYkOIQQWZPgEEJk7f8DH180GVan2CsAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 288x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(4,8))\n",
    "sns.violinplot(x='group', y='value', data=y)#, scale=\"count\", inner=\"stick\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [default]",
   "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": "3.5.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}