05-Pandas/02-pandas-interpolation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "탐색적 데이터 분석 (Exploratory Data Analysis)를 진행하다보면, 매우 흔한 확률로 **결측치가 존재**하는 것을 볼 수 있습니다.\n",
    "\n",
    "우리는 이런 경우 결측치를 **버릴건지(drop), 혹은 채워**주어야 합니다.\n",
    "\n",
    "그러나, 데이터를 함부로 **drop하여 머신러닝 예측을 진행한다는 것은 꽤 risk** 하기도 합니다. 물론 결측치가 매우 적은 경우엔 그냥 drop하는 것이 맞을 수 있습니다.\n",
    "\n",
    "하지만, 결측치가 꽤 많이 존재하는 경우, 혹은 우리가 예측해야할 test set의 feature 데이터에 존재하는 경우에는 결측치를 채워줄 수 밖에 없습니다.\n",
    "\n",
    "결측치를 채워 줄 때 흔히 사용되는 방법은 `0` 이나 `-1`로 단순하게 채워주거나, 조금 더 발전된 방법은 `mean`이나 `median`으로 채워주는 방법입니다. 조금 더 나아간다면, 분류형 컬럼을 기준으로 `groupby`하여 `mean`이나 `median`으로 채워줄 수 있겠네요.\n",
    "\n",
    "하지마, 시계열(Time Series) 데이터나 연속된 수치를 가지는 데이터의 경우에는 우리는 일종의 연속성있는 패턴을 발견할 수 있습니다. 이런 경우 보간(Interpolation)을 통해 **앞,뒤 값을 통하여 유추하여 좀 더 스마트하게 결측치(NaN)를 채워**줄 수 있습니다.\n",
    "\n",
    "대부분, `Pandas`에 이런 유용한 기능이 내장되어 있는 점을 모르는 분들이 많은데, 이번 포스팅에서는 결측치에 대하여 보간(Interpolation) 처리를 해주는 방법에 대하여 알아보도록 하겠습니다."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**참고(Reference) - Pandas 공식 도큐먼트**\n",
    "\n",
    "[pandas.DataFrame.interpolate](https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.interpolate.html)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import scipy.stats as stats\n",
    "import matplotlib.pyplot as plt\n",
    "import math\n",
    "import warnings\n",
    "\n",
    "warnings.filterwarnings('ignore')\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "실습을 위하여 **Normal Distribution을 따르는 샘플을 생성**하도록 하겠습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 평균\n",
    "mu = 0\n",
    "# 분산\n",
    "variance = 1\n",
    "# sigma (Standard Deviation 계산)\n",
    "sigma = math.sqrt(variance)\n",
    "x = np.linspace(mu - 3*sigma, mu + 3*sigma, 100)\n",
    "\n",
    "# 랜덤한 10개의 데이터를 삭제합니다.\n",
    "idx = np.random.choice(len(x), size=10)\n",
    "x[idx] = np.nan\n",
    "\n",
    "# 시각화\n",
    "plt.plot(x, stats.norm.pdf(x, mu, sigma))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "총 10개의 NaN 값을 만들었습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "10"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(x).isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "중간 중간에 NaN값이 포진해 있는 것을 확인할 수 있습니다."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    -3.000000\n",
       "1    -2.939394\n",
       "2    -2.878788\n",
       "3    -2.818182\n",
       "4    -2.757576\n",
       "5    -2.696970\n",
       "6          NaN\n",
       "7    -2.575758\n",
       "8    -2.515152\n",
       "9    -2.454545\n",
       "10   -2.393939\n",
       "11         NaN\n",
       "12   -2.272727\n",
       "13   -2.212121\n",
       "14   -2.151515\n",
       "15   -2.090909\n",
       "16         NaN\n",
       "17   -1.969697\n",
       "18   -1.909091\n",
       "19   -1.848485\n",
       "dtype: float64"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.Series(x).head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 보간 (Interpolation)을 활용한 결측치 대입"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 주요 Hyperparameter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**methodstr, default `linear`**\n",
    "Interpolation technique 지정\n",
    "\n",
    "-`linear`: 색인을 무시하고 값을 동일한 간격으로 처리 (MultiIndexes에서 지원되는 유일한 방법)\n",
    "\n",
    "- `time`: \n",
    "주어진 간격의 길이를 보간하기 위해 매일 더 높은 해상도 데이터를 처리\n",
    "\n",
    "- `index`, `values`: \n",
    "인덱스의 실제 숫자 값을 사용\n",
    "\n",
    "- `pad`: \n",
    "기존 값을 사용하여 NaN 채우기\n",
    "\n",
    "- `nearest`, `zero`, `slinear`, `quadratic`, `cubic`, `spline`, `barycentric`, `polynomial`: \n",
    "scipy.interpolate.interp1d로 전달. 이 방법은 색인의 숫자 값을 사용. `polynomial`과 `spline` 모두 순서 (int)도 지정해야합니다 (예 : df.interpolate (method='polynomial', order=5).\n",
    "\n",
    "- `krogh`, `piecewise_polynomial`, `spline`, `pchip`, `akima`: 비슷한 이름의 SciPy 보간 방법을 둘러싼 Wrapper.\n",
    "\n",
    "- `from_derivatives`: scipy 0.18의 `piecewise_polynomial`보간 방법을 대체하는 scipy.interpolate.BPoly.from_derivatives를 참조\n",
    "\n",
    "**axis{0 or `index`, 1 or `columns`, None}, default None**\n",
    "보간할 축 설정\n",
    "\n",
    "**limitint, optional**\n",
    "채울 최대 연속 NaN 갯수. 0보다 커야함.\n",
    "\n",
    "**inplacebool, default False**\n",
    "데이터 업데이트 (if possible)\n",
    "\n",
    "**limit_direction{`forward`, `backward`, `both`}, default `forward`**\n",
    "limit이 지정되면, 연속 NaN이 `지정된 방향`으로 채워집니다.\n",
    "\n",
    "**limit_area{None, `inside`, `outside`}, default None**\n",
    "limit이 지정되면, 연속된 NaN은 지정된 제한(restriction)으로 채워집니다.\n",
    "\n",
    "- None: No fill restriction.\n",
    "\n",
    "- `inside`: 유효한 값으로 둘러싸인 NaN 만 채 웁니다 (보간).\n",
    "\n",
    "- `outside`: 유효 값을 초과하는 NaN 만 채 웁니다 (외삽).\n",
    "\n",
    "(New in version 0.23.0.)\n",
    "\n",
    "**downcastoptional, `infer` or None, defaults to None**\n",
    "Downcast dtypes if possible."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_inter = pd.Series(x).interpolate()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    -3.000000\n",
       "1    -2.939394\n",
       "2    -2.878788\n",
       "3    -2.818182\n",
       "4    -2.757576\n",
       "5    -2.696970\n",
       "6    -2.636364\n",
       "7    -2.575758\n",
       "8    -2.515152\n",
       "9    -2.454545\n",
       "10   -2.393939\n",
       "11   -2.333333\n",
       "12   -2.272727\n",
       "13   -2.212121\n",
       "14   -2.151515\n",
       "15   -2.090909\n",
       "16   -2.030303\n",
       "17   -1.969697\n",
       "18   -1.909091\n",
       "19   -1.848485\n",
       "dtype: float64"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_inter.head(20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 보간(Interpolation) 이후 채워진 값에 대한 시각화"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD8CAYAAACb4nSYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzt3Xl4VeW5/vHvkxlCCFPCkBkSZpkMAUScQMVqwakKFIE6t6L2+PNUe9pqa21Pq63H1tJTOY6oiCjWUsWiqKDIGEaZAiEDGYAkBAiEzHl+f2TbphHIJiRZe3g+15XLvddeK9z7ktysrP2u9xVVxRhjjH8IcDqAMcaY9mOlb4wxfsRK3xhj/IiVvjHG+BErfWOM8SNW+sYY40es9I0xxo9Y6RtjjB+x0jfGGD8S5HSApnr06KGJiYlOxzDGGK+yadOmElWNam4/jyv9xMRE0tPTnY5hjDFeRURy3dnPLu8YY4wfsdI3xhg/YqVvjDF+xErfGGP8iFulLyKTRSRDRDJF5NGz7HeziKiIpDba9mPXcRkicnVrhDbGGNMyzY7eEZFAYB5wJZAPbBSRpaq6q8l+EcADwPpG2wYD04AhQB9ghYj0V9W61nsLxhhj3OXOmX4akKmqWapaDSwCpp5mv18CTwGVjbZNBRapapWqZgOZru9njDHGAe6M048B8ho9zwfGNN5BREYCcar6vog83OTYdU2OjWlhVmPaXVFZJRtyStl7+CS4lhYNDQ5keGwXRsZ3ITzU4251Meas3PkbK6fZ9s+FdUUkAPgfYM65Htvoe9wN3A0QHx/vRiRj2s7R8mpeXpPD37YWkHvk1D+3i+tv89fLSgcGCMNiI5k1LoFvD+tDUKCNizCez53SzwfiGj2PBQobPY8AhgIrpeGnohewVESmuHEsAKo6H5gPkJqaaiu1G0cUnajk+VVZLFx/gIqaOi7tH8VtYxMYndiNwX06E+wq9ROVNWw+cIyN2aV8tOsQ//HWNn7/0V7uubQf00bH/XM/YzyRqJ69Y0UkCNgLTAQKgI3ADFXdeYb9VwIPq2q6iAwBFtJwHb8P8AmQcrYPclNTU9WmYTDt7R87DvHjd7dTVlnL1BF9+P6l/UjpGdHscfX1yqd7ipi3MpMtB44xPDaS/7l1BH2jOrVDamP+RUQ2qWpqc/s1e6avqrUiMhdYDgQCL6nqThF5AkhX1aVnOXaniCwGdgG1wH02csd4kvKqWn7x950sTs9naExn3r51BMnRzZf91wIChEmDezJxUDTLvjrEf/31K67942p+dt1gpqfFIXK6K5zGOKfZM/32Zmf6pr0Unahk1osbyDh8gh9c1o8HJ/YnJOj8Ls0cOl7Jw29vY3VmCdPT4nny+qEEBljxm7bXamf6xviivNJTzHxxPcUnqnj1e2lc0r/ZGWnd0isyjAW3p/G7jzL488r9lFXW8D+3jDjvf0yMaS1W+sbv7D18gpkvrKeqtp437hzDyPiurfr9AwKEH00eSJeOwfx62R7KKmp4/rYL6RhiP27GeXb6YfxKXukpZvxfw03ji+8Z1+qF39jdl/TjqZuG8WVmCT94YzO1dfVt9mcZ4y4rfeM3jp+qYc7LG6iureONO8cwoJf7H9i21C2j43jy+gtYmVHMz/62A0/7DM34H/t90/iFqto67notnbzSChbckebWcMzWMmNMPPlHT/HnlfuJ7dqR+y5Pbrc/25imrPSNz1NVHnlnOxuyS/nj9JGM7du93TP859UDKDxWwdPLM4jr1pEpw/u0ewZjwC7vGD+wcMMB3ttayP+7sr9jZSsiPHXzcEYnduXHS7aTVXzSkRzGWOkbn7arsIxf/H0Xl/SPcvyySkhQAH+cPpKQoADuW7iFyhq7T9G0Pyt947NOVtUyd+FmunQI5plbhhPgATdJ9Y7swDO3jGD3wTJ++f6u5g8wppVZ6Ruf9dh7O8g5Us4fp4+kR6dQp+P80+UDo7nnkr68sf4AH2w/6HQc42es9I1P+mjnId7dUsD9V6Q48sFtcx6+egDD47rw0/e+ouRkldNxjB+x0jc+5/ipGn763g4G9e7M3Cs8c3hkcGAAT988jPKqOn6+9LQT1hrTJqz0jc958oNdHCmv5umbh3n03Pb9e0Zw/xXJvL/9IP/YccjpOMZPeO5PhDEt8PneYt7elM89l/RlaEyk03Gade9l/RjcuzM/+9sOjp+qcTqO8QNW+sZnVFTX8eN3v6JfVDgPTExxOo5bggMDeOrmYZSWV/PrZbudjmP8gJW+8Rn/u2o/Bccq+PUNFxAWHOh0HLcNjYnkjouTeCs9j615x5yOY3ycW6UvIpNFJENEMkXk0dO8fq+IfCUiW0VktYgMdm1PFJEK1/atIvKX1n4DxkDD7Jl/WbWfKcP7MMYDR+s05/4rkomKCOXxv+2gvt4mZTNtp9nSF5FAYB5wDTAYmP51qTeyUFUvUNURwFPAM41e26+qI1xf97ZWcGMa++X7uwgU4cffGuh0lBaJCAvm0ckD2ZZ/nHc25zsdx/gwd87004BMVc1S1WpgETC18Q6qWtboaThgpyqm3Xy+t5iPdh1m7hXJ9I7s4HScFrthZAyj4rvw1D/2UFZpH+qatuFO6ccAeY2e57u2/RsRuU9E9tNwpv9Ao5eSRGSLiKwSkQnnldaYJmrq6vnF33eS2L0jd05IcjrOeQkIEJ6YOpQj5dX8YcU+p+MYH+VO6Z9uwpJvnMmr6jxV7Qc8AvzUtfkgEK+qI4GHgIUi0vkbf4DI3SKSLiLpxcXF7qc3fu+tjXnsLy7nJ9cOJjTIez68PZOhMZHccmEcC9bmkFd6yuk4xge5U/r5QFyj57FA4Vn2XwRcD6CqVap6xPV4E7Af6N/0AFWdr6qpqpoaFdU6C1Qb31deVcuzK/YxOrErkwZFOx2n1fzHlf0JDBB+91GG01GMD3Kn9DcCKSKSJCIhwDRgaeMdRKTxoOhrgX2u7VGuD4IRkb5ACpDVGsGNeXF1NiUnq3j0mkGIOD+DZmvpFRnG7eOT+NvWQnYUHHc6jvExzZa+qtYCc4HlwG5gsaruFJEnRGSKa7e5IrJTRLbScBlntmv7JcB2EdkGvAPcq6qlrf4ujN85crKK51ft5+ohPbkwoe0WN3fKPZf2o0vHYH77jz1ORzE+xq3lElV1GbCsybbHGj1+8AzHLQGWnE9AY07nuU8zqaip4z+v9s4hms2J7BDM3MuTefKD3azeV8LFKT2cjmR8hN2Ra7xOXukp3lify62j40iO7uR0nDYzc2wCMV068NTyPajaKGjTOqz0jdeZ91kmgnjN/DotFRYcyIMTU9ief5xPdhc5Hcf4CCt941XySk/xzqZ8pqfFefWNWO66YVQM8d068uwne+1s37QKK33jVf70aSYBAcL3L/PMxVFaW3BgAPdfkcyOgjJW2Nm+aQVW+sZrHDhyinc25zMjLZ5ekWFOx2k3N4yMIaF7R55dYWf75vxZ6Ruv8dyn+wgKEL5/WT+no7SroMAA7r8ihZ2FZXy067DTcYyXs9I3XuHAkVO8u6WAGWPi6dnZf87yv3b9iD4k9QjnDyv22dm+OS9W+sYr/OXz/QSKcO+l/nWW/7WgwAB+cFk/dh0sY2WGzU9lWs5K33i8orJK3knP5+bUWL88y//a9SNjiOnSgXmfZTodxXgxK33j8V5YnU1tfT33XuKfZ/lfCw4M4O5L+pKee5QN2TabiWkZK33j0Y6dquaNdbl8e3gf4rt3dDqO425JjaN7eIid7ZsWs9I3Hu3VNbmUV9f53YidM+kQEsjtFyexam+xzcBpWsRK33is8qpaXl6TzaRB0Qzs9Y21d/zWbeMSiAgN4s8r7WzfnDsrfeOx3tqYx7FTNX5z9627OocFM3NcAh/uOEROSbnTcYyXsdI3Hqm2rp4XV2czOrGrT86Xf76+d1EiwQEBvLDa1iQy58ZK33ikZTsOUXCsgrsm9HU6ikeK7hzG9SP78HZ6PkdOVjkdx3gRt0pfRCaLSIaIZIrIo6d5/V4R+UpEtorIahEZ3Oi1H7uOyxCRq1szvPFNqsr8z/fTt0c4kwb1dDqOx7prQl+qaut5fd0Bp6MYL9Js6bvWuJ0HXAMMBqY3LnWXhap6gaqOAJ4CnnEdO5iGNXWHAJOBP3+9Zq4xZ7Iuq5QdBWXcOaEvAQG+s/Zta0vpGcEVA6NZsDaHypo6p+MYL+HOmX4akKmqWapaDSwCpjbeQVXLGj0NB76eHGQqsEhVq1Q1G8h0fT9jzmj+5/vpHh7CjaNinI7i8e6a0Jcj5dUs2ZzvdBTjJdwp/Rggr9HzfNe2fyMi94nIfhrO9B84l2ON+drewyf4LKOY2RclEhZsvxQ2Z2zfbgyLjeSFL7Kpr7eJ2Ezz3Cn90/1+/Y2/Xao6T1X7AY8APz2XY0XkbhFJF5H04mKbTMqfvbQ6m9CgAGaOTXA6ilcQEe6c0JfsknI+3WOLrJjmuVP6+UBco+exQOFZ9l8EXH8ux6rqfFVNVdXUqKgoNyIZX3TkZBXvbingxlGxdAsPcTqO17hmaC96R4bx0pfZTkcxXsCd0t8IpIhIkoiE0PDB7NLGO4hI4xWqrwX2uR4vBaaJSKiIJAEpwIbzj2180cL1B6iuref28YlOR/EqwYEBzL4okTX7j7CrsKz5A4xfa7b0VbUWmAssB3YDi1V1p4g8ISJTXLvNFZGdIrIVeAiY7Tp2J7AY2AX8A7hPVW2YgfmG6tp6FqzL5ZL+UaT0jHA6jteZPjqeDsGBvGxn+6YZQe7spKrLgGVNtj3W6PGDZzn2V8CvWhrQ+If3txdSfKKKp29OdDqKV4rsGMzNF8by1sY8fjR5IFERoU5HMh7K7sg1jlNVXlydTXJ0Jy7tb5/ptNSc8YlU19Xzxvpcp6MYD2albxy3MecoOwvL+N74RETsZqyW6hfViSsGRvP6ulyqau0qqjk9K33juJe/zCayQzA3jox1OorXu318EiUnq3l/20GnoxgPZaVvHFVwrIKPdh1mWlocHULsZqzzNT65O8nRnXhlTQ6qdrOW+SYrfeOo19floqrcZjdjtQoRYc5FiXxVcJzNB445Hcd4ICt945jKmjre3HCAqwb3IrarrX/bWm4YGUNEWBCvrMlxOorxQFb6xjFLtxZy7FQNc+xmrFYVHhrEralxfPjVQQ6XVTodx3gYK33jCFXl5TU5DOwVwZikbk7H8TmzxiVSp8ob62z4pvl3VvrGERuyS9l9sIw5F9kwzbYQ370jEwdG88b6AzZ80/wbK33jiAVrc4nsEMzUETbTdluZc1ESR8qrWfaVDd80/2Klb9rdweMV/GPnIaaNtmGabWl8cnf6RoXzyhq7xGP+xUrftLuF6w9Qr2pz5rcxEWH2uES25R1ja54N3zQNrPRNu6qqbRimOXFgNHHdbJhmW7txVAzhIYEsWJvjdBTjIaz0Tbv68KtDlJysZta4RKej+IWIsGBuujCW97cd5MjJKqfjGA9gpW/a1atrc+jbI5yLk3s4HcVvzBqXQHVdPYs25jW/s/F5Vvqm3WzPP8aWA8e4bVwCAQE2TLO9JEdHMD65O2+sy6W2rt7pOMZhbpW+iEwWkQwRyRSRR0/z+kMisktEtovIJyKS0Oi1OhHZ6vpa2vRY4z8WrM2lY0ggN11os2m2t1njEik8XsmK3bZ4ur9rtvRFJBCYB1wDDAami8jgJrttAVJVdRjwDvBUo9cqVHWE62sKxi8dLa9m6bZCbhgZQ+ewYKfj+J2JA6PpExnGa+tynI5iHObOmX4akKmqWapaDSwCpjbeQVU/U9VTrqfrADuVM//mrfQ8qmvr7QNchwQFBvDdsQl8mXmEzKITTscxDnKn9GOAxp8A5bu2nckdwIeNnoeJSLqIrBOR61uQ0Xi5unrl9XW5pCV1Y0AvW/TcKdNGxxESGMBra+1mLX/mTumf7hO3067OICIzgVTg6Uab41U1FZgBPCsi/U5z3N2ufxjSi4uL3YhkvMnKjCLyj1Ywa5zdjOWk7p1CuXZYb5ZsLuBkVa3TcYxD3Cn9fCCu0fNYoLDpTiIyCfgJMEVV/zkgWFULXf/NAlYCI5seq6rzVTVVVVOjomxhbF+zYG0u0RGhXD2kl9NR/N5t4xI4WVXLX7cUOB3FOMSd0t8IpIhIkoiEANOAfxuFIyIjgedpKPyiRtu7ikio63EPYDywq7XCG8+XU1LOqr3FzBgTT3CgjRB22si4LgyN6cxra205RX/V7E+hqtYCc4HlwG5gsaruFJEnROTr0ThPA52At5sMzRwEpIvINuAz4DeqaqXvR15fl0tQgDAjLd7pKIaG+XhmjUtk7+GTrMsqdTqOcUCQOzup6jJgWZNtjzV6POkMx60BLjifgMZ7VVTXsTg9j6uH9iK6c5jTcYzLlOF9+PWy3by2Lodx/bo7Hce0M/t927SZv20toKyyllk2m6ZHCQsO5JbUOJbvPMyh47acor+x0jdtQlVZsDaXgb0iSLPlED3OzDEJ1KuycMMBp6OYdmalb9rE5gNH2XWwjNvGJdhyiB4ovntHLusfxZsbDlBda/Px+BMrfdMmXl2TS0RoENfbcogea9ZFiRSfqOIfOw85HcW0Iyt90+qKT1Tx4Y6D3HRhLOGhbo0VMA64NCWKhO4deW1tjtNRTDuy0jetbtGGA9TUKbfZHbgeLSBAmDkmgY05R9lVWOZ0HNNOrPRNq6qtq+eN9QeYkNKDflGdnI5jmvGd1FhCgwJ4bZ3Nx+MvrPRNq/p412EOlVXabJpeokvHEK4fEcN7Wwo4fqrG6TimHVjpm1b16tocYrp04IqB0U5HMW66bVwCFTV1vL3JllP0B1b6ptVkHDrBuqxSbhuXQKAth+g1hsZEkprQldfW5VJfb/Px+DorfdNqFqzNITQogFtT45rd13iWWRclknvkFKv22dTmvs5K37SKssoa/rqlgCnD+9A1PMTpOOYcTR7Si6iIUBasyXE6imljVvqmVSzZlM+p6jpmX5TodBTTAiFBAcxIi2fl3mJySsqdjmPakJW+OW/19cpra3MZGd+FoTGRTscxLTRjTDyBIjZ808dZ6Zvz9vm+YrJKypljZ/lerWfnMK65oDeL0/Mot+UUfZaVvjlvr6zJISoilGuG9nY6ijlPcy5K4ERlLe/acoo+y63SF5HJIpIhIpki8uhpXn9IRHaJyHYR+UREEhq9NltE9rm+ZrdmeOO87JJyVmYU890x8YQE2TmEtxsV35ULYiJZsMaWU/RVzf6UikggMA+4BhgMTBeRwU122wKkquow4B3gKdex3YDHgTFAGvC4iHRtvfjGaa+uySE4UJgxxpZD9AUiwuyLEtlXdJI1+484Hce0AXdOzdKATFXNUtVqYBEwtfEOqvqZqp5yPV0HxLoeXw18rKqlqnoU+BiY3DrRjdNOVtXyzqZ8rr2gN9ERthyir7huWG+6h4fw8pc5TkcxbcCd0o8BGt+fne/adiZ3AB+28FjjRZZsyudkVa0N0/QxYcGBTE+L55M9h8krPdX8AcaruFP6p7uf/rQX+0RkJpAKPH0ux4rI3SKSLiLpxcV2R6A3qK9XXl2bw/DYSEbG2xU7XzNzbAKBIixYm+N0FNPK3Cn9fKDxffWxQGHTnURkEvATYIqqVp3Lsao6X1VTVTU1KirK3ezGQav2FZNVXM7tFyc5HcW0gV6RYUwe2otFG234pq9xp/Q3AikikiQiIcA0YGnjHURkJPA8DYVf1Oil5cBVItLV9QHuVa5txsu9tDqbaBum6dNuvziJE5W1LNmc73QU04qaLX1VrQXm0lDWu4HFqrpTRJ4QkSmu3Z4GOgFvi8hWEVnqOrYU+CUN/3BsBJ5wbTNebN/hE3yxr4RZ4xJsmKYPGxXflRFxXXj5yxybfdOHuLWAqaouA5Y12fZYo8eTznLsS8BLLQ1oPM/Laxpm05yeZsM0fd33xify4KKtrNxbxBUDezodx7QCO00z5+TYqWre3ZzPDSNj6N4p1Ok4po1964Le9OwcasM3fYiVvjknb27Io7KmnjnjE52OYtpBcGAAs8Yl8sW+EvYePuF0HNMKrPSN22rq6lmwNofxyd0Z2Kuz03FMO5mRFk9oUAAvrc52OoppBVb6xm0fbD/IweOV3GHDNP1K1/AQbhwVy7tbCig5WdX8AcajWekbt6gqL6zOol9UOJf1t0XP/c0dFydRXVvPa2ttrn1vZ6Vv3LIuq5QdBWXcOaEvAbboud9Jju7ExIHRvLYul8qaOqfjmPNgpW/c8sIXWXQPD+GGkTZ1kr+6Y0ISpeXV/NXm2vdqVvqmWZlFJ/lkTxEzxyYQFhzodBzjkHF9uzOkT2de+CLLbtbyYlb6plkvrs4mJCiA28YlNL+z8Vkiwl0T+rK/uJyVe4uaP8B4JCt9c1YlJ6t4d3M+N46MoYfdjOX3rh3Wm16dw5j/eZbTUUwLWembs3p1TQ7VdfXcOaGv01GMBwgODOD2ixNZl1XKtrxjTscxLWClb86ovKqWBWtzuXJQT5KjOzkdx3iI6WnxRIQF8fzn+52OYlrASt+c0Vsb8zheUcM9l/ZzOorxIBFhwcwcm8CHOw6RXVLudBxzjqz0zWnV1NXz4ups0hK7cWGCrYxl/t33xicSHBDA/31h1/a9jZW+Oa33txdScKyCey61a/nmm6Ijwrjpwhje2ZRP0YlKp+OYc2Clb75BVXl+VRb9e3bi8gE25YI5vbsm9KWmrp5XbNplr+JW6YvIZBHJEJFMEXn0NK9fIiKbRaRWRG5u8lqdazWtf66oZTzbZxlF7Dl0gnsu6WdTLpgz6hvViclDevHaulzKKmucjmPc1Gzpi0ggMA+4BhgMTBeRwU12OwDMARae5ltUqOoI19eU07xuPIiq8tynmcR27cCUEX2cjmM83A8uS+ZEZa1NxOZF3DnTTwMyVTVLVauBRcDUxjuoao6qbgfq2yCjaUdr9x9hy4Fj3HtpP4ID7eqfObsLYiO5bEAUL63OpqLaJmLzBu78VMcAeY2e57u2uStMRNJFZJ2IXH9O6Uy7+9NnmURHhHLzhbFORzFeYu7lyRwpr+bNDQecjmLc4E7pn+6i7rnMthSvqqnADOBZEfnGoG8Rudv1D0N6cXHxOXxr05o2HzjKmv1HuPuSvjaxmnFbamI3xiR1Y/7nWVTV2tm+p3On9POBuEbPY4FCd/8AVS10/TcLWAmMPM0+81U1VVVTo6Ki3P3WppXN+zSTrh2DmTEm3ukoxsvMvSKZQ2WVvLvZpl32dO6U/kYgRUSSRCQEmAa4NQpHRLqKSKjrcQ9gPLCrpWFN29lRcJxP9hRx+/gkOoYEOR3HeJmLk3swPK4Lf16ZSU2dfbTnyZotfVWtBeYCy4HdwGJV3SkiT4jIFAARGS0i+cB3gOdFZKfr8EFAuohsAz4DfqOqVvoe6A+f7KNzWBCzxyc6HcV4IRHhwYnJ5JVW8Fc72/dobp3SqeoyYFmTbY81eryRhss+TY9bA1xwnhlNG9tRcJyPdx3moSv70zks2Ok4xktdPiCa4bGRPPfZPm4YFWOjvzyU/V8xPLtiL5EdgpljZ/nmPIgIP5zUn7zSCpZsync6jjkDK30/91X+cVbsLuLOi5PsLN+ct8sGRDE8NpI/fZZJda1d2/dEVvp+zs7yTWv6+mw//2gFSzbb2b4nstL3Y1vzjvHJniLumpBEhJ3lm1Zy2YAohsd14U+fZtq4fQ9kpe/Hnl6+h27hIcy+KNHpKMaHiAj/78r+FByrYOF6u0vX01jp+6kvM0v4MvMIP7isn53lm1Y3IaUHY/t240+fZlJeVet0HNOIlb4fUlWeWp5Bn8gwZo5NcDqO8UEiwo8mD+RIeTUvrc52Oo5pxErfDy3feZhtecd4cFKKzbFj2syo+K5MGtST+Z9ncbS82uk4xsVK38/U1Su//yiDvlHh3DTKZtI0bes/rx7Ayepa/rJqv9NRjIuVvp9ZsjmffUUnefiqAQTZHZOmjQ3oFcENI2J4ZU0OhccqnI5jsNL3K6eqa/nd8gxGxHXhmqG9nI5j/MRDV/VHgd8tz3A6isFK36/M/zyLohNV/PTaQYjY2remfcR27cjt45N4d0sBX+UfdzqO37PS9xNFZZU8vyqLa4b2IjWxm9NxjJ/5weX96BYewq+W7UL1XNZgMq3NSt9P/P6jvdTW1/PI5IFORzF+qHNYMD+clMK6rFJW7C5yOo5fs9L3A7sPlrF4Ux63jU0ksUe403GMn5qeFk/fqHD++8PdttCKg6z0fZyq8vOlO4nsEMwDE5OdjmP8WHBgAD/51iCyist5dU2O03H8llulLyKTRSRDRDJF5NHTvH6JiGwWkVoRubnJa7NFZJ/ra3ZrBTfueX/7QdZnl/LwVQPo0jHE6TjGz10xMJrLB0Tx7Ip9FJ2odDqOX2q29EUkEJgHXAMMBqaLyOAmux0A5gALmxzbDXgcGAOkAY+LSNfzj23cUV5Vy68+2M2QPp2ZnmaLnRvniQiPfXsI1bX1/ObDPU7H8UvunOmnAZmqmqWq1cAiYGrjHVQ1R1W3A00v1F0NfKyqpap6FPgYmNwKuY0b5n2WyaGySp6YOoTAABuiaTxDUo9w7piQxLubC9iUW+p0HL/jTunHAHmNnue7trnjfI415yG7pJwXvsjmxpExXJhgQzSNZ5l7eTK9Oofx+NKd1NXbEM725E7pn+4U0d3/S24dKyJ3i0i6iKQXFxe7+a3NmagqP3tvByFBATx6jQ3RNJ4nPDSIn1w7iB0FZby2NsfpOH7FndLPB+IaPY8FCt38/m4dq6rzVTVVVVOjoqLc/NbmTP66pYDVmSU8MnkA0Z3DnI5jzGldN6w3l/SP4unlGTYvTztyp/Q3AikikiQiIcA0YKmb3385cJWIdHV9gHuVa5tpI6Xl1Tz5wW5Gxnfhu2NsrnzjuUSEX10/lDpVHl+60+k4fqPZ0lfVWmAuDWW9G1isqjtF5AkRmQIgIqNFJB/4DvC8iOx0HVsK/JKGfzg2Ak+4tpk28utluymrqOG/b7yAAPvw1ni4uG4d+Y9J/fl412H+seOQ03H8gnjaPBipqamanp7udAyvtCazhBkvrOcHl/XjRzbdgvEStXX1TPnTlxxyM85UAAAPdUlEQVQpr+Ljhy6lsy3f2SIisklVU5vbz+7I9REnq2r50ZLtJHbvyAMTU5yOY4zbggID+M1NF1B8ooon39/ldByfZ6XvI369bDcFxyr43XeG2xKIxusMi+3C9y/rx+L0fD7dc9jpOD7NSt8HfL63mIXrD3DXhL42bbLxWg9MTGFgrwgeXfIVx07ZmrptxUrfyx2vqOGRJdtJju7EQ1f2dzqOMS0WGhTI728ZTml5NT+30Txtxkrfy/186U6KTlTxe7usY3zAkD6R3H9FCu9tLeSD7QedjuOTrPS92Lub8/nrlgLuvyKZ4XFdnI5jTKv4weX9GBHXhUff3U5e6Smn4/gcK30vlV1Szk/f20FaUjfuv8JG6xjfERwYwHPTR4LCg4u22IIrrcxK3wtV19Zz/5ubCQkK4A/TRtgMmsbnxHXryK9vvIDNB47xhxX7nI7jU6z0vdBvPtzDjoIynrppGL0jOzgdx5g28e3hfbglNZZ5KzNZva/E6Tg+w0rfyyzdVshLX2Yz56JErhrSy+k4xrSpn08ZQkp0J+5/czP5R+36fmuw0vciew6V8cg72xmd2JX/+tYgp+MY0+Y6hgTxl5kXUlunfP/1zVTW1DkdyetZ6XuJ4xU13PPaJiLCgpg3YxQhQfa/zviHvlGdeObWEXxVcJyfvbcDT5svzNtYc3iB2rp6Hly0hcJjFfzvzFE2R77xO1cO7skDVyTz9qZ8FqzNdTqOV7PS93CqyhPv72JlRjG/mDLUlj40fuvBSf2ZNCiaX/x9J5/tKXI6jtey0vdwL3+Zw4K1udx9SV9mjIl3Oo4xjgkMEP4wbSSDendm7sLN7CosczqSV7LS92Ardh3mlx/s4uohPXnU5sc3hvDQIF6cPZqIsGDueHUjh8sqnY7kddwqfRGZLCIZIpIpIo+e5vVQEXnL9fp6EUl0bU8UkQoR2er6+kvrxvddG7JLmfvmZi6IieTZW0faKljGuPSKDOOlOaMpq6hh1osbbEbOc9Rs6YtIIDAPuAYYDEwXkcFNdrsDOKqqycD/AL9t9Np+VR3h+rq3lXL7tB0Fx7njlY306dKBl+eMpkOITaRmTGOD+3Rm/qxUskvKmf3yRk5W1TodyWu4c6afBmSqapaqVgOLgKlN9pkKvOp6/A4wUUTs1LQFMotOMuulDXTuEMzrd4yhe6dQpyMZ45HGJ/fguRkj2VFwnLsXpNsYfje5U/oxQF6j5/mubafdx7WQ+nGgu+u1JBHZIiKrRGTCeeb1aVnFJ5n5wnoCRHj9zjH06WJTLBhzNlcP6cXTNw9jzf4j3Pv6Jit+N7hT+qc7Y296d8SZ9jkIxKvqSOAhYKGIdP7GHyByt4iki0h6cXGxG5F8z97DJ7jl+XXU1NXz+p1pJPUIdzqSMV7hxlGx/PeNF7BqbzF3vppORbUV/9m4U/r5QFyj57FA4Zn2EZEgIBIoVdUqVT0CoKqbgP3AN5Z3UtX5qpqqqqlRUVHn/i683K7CMqbNX0eAwFv3jGVgr2/8u2iMOYvpafE8ddMwvtxfwpyXN1Bu1/jPyJ3S3wikiEiSiIQA04ClTfZZCsx2Pb4Z+FRVVUSiXB8EIyJ9gRQgq3Wi+4YN2aVMm7+W0KAA3rpnHMnREU5HMsYrfSc1jmdvHUF67lFmvLCekpNVTkfySM2Wvusa/VxgObAbWKyqO0XkCRGZ4trtRaC7iGTScBnn62GdlwDbRWQbDR/w3quqpa39JrzVB9sPMvOF9fSICGXxPePsko4x52nqiBj+97uj2HOwjJv+dw05JeVOR/I44mmTF6Wmpmp6errTMdqUqvLi6mye/GA3qQld+b9ZqXQND3E6ljE+Y1PuUe58dSMiwguzUxkV39XpSG1ORDapampz+9kdue2ssqaOh9/ezpMf7Oaaob14/c4xVvjGtLILE7qy5PsX0Sk0iGnz17F4Y17zB/kJK/12VHCsgu/8ZS1LNufz4MQU5s0YRViw3XhlTFvoG9WJ9+4bT1piN360ZDs/fe8rqmttvd0gpwP4i0/3HObht7dTU1vP/81K5crBPZ2OZIzP6xYewivfG83TH2Xw/KosdhaW8cdpI4nr1tHpaI6xM/02VllTx+N/28Htr6TTs3MY780db4VvTDsKCgzgx9cMYt6MUWQWneRbf/iC97YUOB3LMXam34a25x/jP9/eTsbhE9xxcRI/mjyA0CC7nGOME64d1pvhcZH8x1tb+eFbW/l0TxE/nzKEbn72mZqVfhs4VV3LMx/t5aUvs4mKCOXV29O4tL//3XRmjKeJ7dqRN+8ay59X7ue5T/exOrOEx64bzNQRffCX6cJsyGYrUlU+2nWYJz/YRV5pBTPGxPPoNQPpHBbsdDRjTBMZh07wyJLtbM07xiX9o3jsusEkR3dyOlaLuTtk00q/lewqLOOX7+9ibdYRkqM78avrhzKmb/fmDzTGOKauXlmwNodnPtrLqZo6bhubwA8npdClo/dd8rHSbyfZJeX88ZN9/G1rAZ07BPPQlf2ZkRZPUKB9Rm6Mtyg5WcUzH+9l0YYDdAoN4q4JfZkzPpEIL/ot3Uq/jWUWneQvq/bz7uZ8QoICmDUukfsuSyayo/f8JTHG/LvdB8v4/UcZrNhdRJeOwdw1oS8zxyYQ2cHzf66t9NuAqrIuq5QXvsjikz1FhAYFMHNsAvdc2pfoiDCn4xljWsm2vGM8u2Ivn2UUEx4SyC2j47h9fJJHj++30m9FR8ureXdLAW9tPMDewyfpFh7CbWMTuG1cAj1sZStjfNaOguO8uDqbv28rpE6VS/tHMW10PBMHRRPsYZdwrfTPU2VNHZ/uKeLv2wr5ZHcR1XX1DI/rwoy0OKaOiLHpE4zxI4eOV/LG+lwWp+dxuKyKHp1CuW5Yb749vA+j4rt4xHBPK/0WOHaqmpUZxazYfZjP9hRRXl1Hj04hXDesD7eOjmNQb1vcxBh/VltXz6q9xSxOz+OzjGKqa+uJ6dKBq4b0ZOLAnqQldSMkyJnfAKz03VBVW8eWA8dYk1nCl/uPsDXvGHX1So9OoUwaFM23h/dhTFI3G4ljjPmGssoaPt55mA++OsiXmSVU1dYTERrE2H7dGd+vO+OTe5Ac3andfguw0m9CVck/WsGOguNsyTvG5tyjbC84TnVtPQECF8R2YUJyDyYOimZ4bBcCApz/dc0Y4x1OVdfyZeYRPt1zmNWZJeSVVgDQtWMwo+K7MiqhK8NjuzCkT+c2m0q9VUtfRCYDfwACgRdU9TdNXg8FFgAXAkeAW1U1x/Xaj4E7gDrgAVVdfrY/63xLX1U5XFZFdkk5mcUn2Xf4BHsPn2BXYRlllQ3rZoYEBjA0pjMXJnRldGI3xvTt7hVDsowx3iGv9BRr9x8hPbeUTblH2V/8rxW8Yrp0YFDvCFJ6RtC/Zyf6RXUisUf4ed+5727pNzv3jmuN23nAlTQsgL5RRJaq6q5Gu90BHFXVZBGZBvwWuFVEBtOwpu4QoA+wQkT6q2qrL1dfVFbJrJc2kHvkFBU1//r2nUKDSOnZieuG92FIn84M7RPJgF4R9kGsMabNxHXrSFy3jtwyOg5o+LxwR0EZOwuPs6OwjIxDZazaW0xN3b9OuruFhzA+uQfPTR/ZptncmXAtDchU1SwAEVkETAUal/5U4Oeux+8Af5KGC1lTgUWqWgVku9bQTQPWtk78f+nSMYTYrh0Yn9yDxB7hJHbvSHJ0J3p1DvOIT9aNMf6rS8cQLk7pwcUpPf65raaunpyScrJKysk9Uk52ySm6hbf9FQd3Sj8GaLzWWD4w5kz7qGqtiBwHuru2r2tybEyL055FSFAAL8we3Rbf2hhjWl1wYAApPRsu87Qnd4alnO40uekHAWfax51jEZG7RSRdRNKLi4vdiGSMMaYl3Cn9fCCu0fNYoPBM+4hIEBAJlLp5LKo6X1VTVTU1KsrmnTfGmLbiTulvBFJEJElEQmj4YHZpk32WArNdj28GPtWGYUFLgWkiEioiSUAKsKF1ohtjjDlXzV7Td12jnwssp2HI5kuqulNEngDSVXUp8CLwmuuD2lIa/mHAtd9iGj70rQXua4uRO8YYY9zjNzdnGWOML3N3nL7NL2CMMX7ESt8YY/yIlb4xxvgRj7umLyLFQO55fIseQEkrxXGSr7wPsPfiqXzlvfjK+4Dzey8JqtrsmHePK/3zJSLp7nyY4el85X2AvRdP5SvvxVfeB7TPe7HLO8YY40es9I0xxo/4YunPdzpAK/GV9wH2XjyVr7wXX3kf0A7vxeeu6RtjjDkzXzzTN8YYcwY+V/oi8ksR2S4iW0XkIxHp43SmlhKRp0Vkj+v9/FVEujidqaVE5DsislNE6kXE60ZaiMhkEckQkUwRedTpPOdDRF4SkSIR2eF0lvMhInEi8pmI7Hb93XrQ6UwtJSJhIrJBRLa53ssv2uzP8rXLOyLSWVXLXI8fAAar6r0Ox2oREbmKhhlLa0XktwCq+ojDsVpERAYB9cDzwMOq6jUTLLmWDN1LoyVDgelNlgz1GiJyCXASWKCqQ53O01Ii0hvoraqbRSQC2ARc743/X1wrDYar6kkRCQZWAw+q6rpmDj1nPnem/3Xhu4RzmkVbvIWqfqSqta6n62hYj8ArqepuVc1wOkcL/XPJUFWtBr5eMtQrqernNMyG69VU9aCqbnY9PgHspo1W5mtr2uCk62mw66tNusvnSh9ARH4lInnAd4HHnM7TSm4HPnQ6hJ863ZKhXlkuvkpEEoGRwHpnk7SciASKyFagCPhYVdvkvXhl6YvIChHZcZqvqQCq+hNVjQPeAOY6m/bsmnsvrn1+QsN6BG84l7R57rwXL+XWsp/GGSLSCVgC/LDJb/peRVXrVHUEDb/Rp4lIm1x6c2dhdI+jqpPc3HUh8AHweBvGOS/NvRcRmQ1cB0xUD/8A5hz+v3gbt5b9NO3Pdf17CfCGqr7rdJ7WoKrHRGQlMBlo9Q/bvfJM/2xEJKXR0ynAHqeynC8RmQw8AkxR1VNO5/Fj7iwZatqZ68PPF4HdqvqM03nOh4hEfT06T0Q6AJNoo+7yxdE7S4ABNIwUyQXuVdUCZ1O1jGv5yVDgiGvTOi8eiXQD8BwQBRwDtqrq1c6mcp+IfAt4ln8tGforhyO1mIi8CVxGw4yOh4HHVfVFR0O1gIhcDHwBfEXDzzvAf6nqMudStYyIDANepeHvVwCwWFWfaJM/y9dK3xhjzJn53OUdY4wxZ2alb4wxfsRK3xhj/IiVvjHG+BErfWOM8SNW+sYY40es9I0xxo9Y6RtjjB/5/+DyTToHTL9MAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(x_inter, stats.norm.pdf(x_inter, mu, sigma))\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Environment (conda_tensorflow_p36)",
   "language": "python",
   "name": "conda_tensorflow_p36"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}