ch03_회귀_알고리즘과_모델_규제.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "ch03 회귀 알고리즘과 모델 규제.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyPOHG9blVg1eGsZ140JANpa"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "Di1b2KqeoAEs"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mnc7qnXgE_Qe"
      },
      "source": [
        "# Chapter3. 회귀 알고리즘과 모델 규제"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vt1Q33cFFLQh"
      },
      "source": [
        "## 03-1. k-최근접 이웃 회귀"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 313
        },
        "id": "M6-KYLyOFraQ",
        "outputId": "be2c2277-7649-4a6a-ca78-83909d8ef82a"
      },
      "source": [
        "import numpy as np\n",
        "perch_length = np.array([8.4, 13.7, 15.0, 16.2, 17.4, 18.0, 18.7, 19.0, 19.6, 20.0, 21.0,\n",
        "       21.0, 21.0, 21.3, 22.0, 22.0, 22.0, 22.0, 22.0, 22.5, 22.5, 22.7,\n",
        "       23.0, 23.5, 24.0, 24.0, 24.6, 25.0, 25.6, 26.5, 27.3, 27.5, 27.5,\n",
        "       27.5, 28.0, 28.7, 30.0, 32.8, 34.5, 35.0, 36.5, 36.0, 37.0, 37.0,\n",
        "       39.0, 39.0, 39.0, 40.0, 40.0, 40.0, 40.0, 42.0, 43.0, 43.0, 43.5,\n",
        "       44.0])\n",
        "perch_weight = np.array([5.9, 32.0, 40.0, 51.5, 70.0, 100.0, 78.0, 80.0, 85.0, 85.0, 110.0,\n",
        "       115.0, 125.0, 130.0, 120.0, 120.0, 130.0, 135.0, 110.0, 130.0,\n",
        "       150.0, 145.0, 150.0, 170.0, 225.0, 145.0, 188.0, 180.0, 197.0,\n",
        "       218.0, 300.0, 260.0, 265.0, 250.0, 250.0, 300.0, 320.0, 514.0,\n",
        "       556.0, 840.0, 685.0, 700.0, 700.0, 690.0, 900.0, 650.0, 820.0,\n",
        "       850.0, 900.0, 1015.0, 820.0, 1100.0, 1000.0, 1100.0, 1000.0,\n",
        "       1000.0])\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "plt.scatter(perch_length, perch_weight)\n",
        "plt.xlabel('length')\n",
        "plt.ylabel('weight')\n",
        "plt.show()\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "train_input, test_input, train_target, test_target = train_test_split(perch_length, perch_weight, random_state=42)\n",
        "\n",
        "print(train_input.shape, test_input.shape) # 1차원 배열\n",
        "\n",
        "# 사이킷런에 사용할 훈련 세트는 2차원 배열이여야 함. -> reshape(-1, 1)\n",
        "train_input = train_input.reshape(-1, 1)\n",
        "test_input = test_input.reshape(-1, 1)\n",
        "\n",
        "print(train_input.shape, test_input.shape)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEGCAYAAACUzrmNAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAbRklEQVR4nO3df5CcdZ3g8feHMOog6KDkEAa4UErFUliJjMpe9rYUToI/FrLoorvqopsrdqv0FnUvS7CsQ/e8I5pVZOu2qENQsWQVQS6wam3kSCjvPOVMSDAKpmQVJCNCXBhklzkYwuf+6KdDz6R7np6Z7n66Z96vqq7p5/s83f3hKdKf/v6OzESSpNkcUnUAkqT+Z7KQJJUyWUiSSpksJEmlTBaSpFKHVh1ANxx11FG5YsWKqsOQpIGyY8eOX2fm8mbnFmWyWLFiBdu3b686DEkaKBFxf6tzNkNJkkqZLCRJpUwWkqRSJgtJUimThSSp1KIcDSVJg2TzznE2bdnDLycmOXZkmPVrVrJ21Wjb53vBZCFJFdq8c5xLbtrN5NR+AMYnJrnkpt0ArF01Wnq+V2yGkqQKbdqy50AiqJuc2s+mLXvaOt8rJgtJqtAvJyZnLS873ysmC0mq0LEjw7OWl53vFZOFJFVo/ZqVDA8tm1Y2PLSM9WtWtnW+V+zglqQK1TupW412KjvfK7EY9+AeGxtLFxKUpLmJiB2ZOdbsnM1QkqRSJgtJUimThSSplMlCklTKZCFJKmWykCSVMllIkkqZLCRJpZzBLUkL0A97TfSCyUKS5qlf9proha4li4j4PPBW4OHMPLkoexFwPbACuA84PzMfjYgArgDeDDwBvDcz7yxecwHw0eJtP5GZ13YrZkmai9n2muhksmhWe6l/fr3sDS9fzraf7OtaDaebNYsvAv8N+FJD2QbgtszcGBEbiuOLgTcBJxWP1wFXAq8rksulwBiQwI6IuCUzH+1i3JLUll7sNdGs9rL+hrsgYGp/Hij78vd/ceA13ajhdK2DOzO/Azwyo/hcoF4zuBZY21D+paz5PjASEccAa4BbM/ORIkHcCpzdrZglaS56sddEs9rL1DN5IFG00und9Ho9GurozHyweP4r4Oji+SjwQMN1e4uyVuUHiYgLI2J7RGzft29fZ6OWpCZ6sdfEQmopnazhVDZ0Nmtro3dsffTMvCozxzJzbPny5Z16W0lqae2qUS477xRGR4YJYHRkmMvOO6WjfQULqaV0sobT69FQD0XEMZn5YNHM9HBRPg4c33DdcUXZOPD6GeW39yBOSWrL2lWjXR35tH7Nyml9FgBDh8S0PotmOl3D6XXN4hbgguL5BcDNDeV/HDWnA48VzVVbgLMi4siIOBI4qyiTpCWhWe1l0x+8ik1vf9W0sneffkJXazjdHDr7FWq1gqMiYi+1UU0bga9FxDrgfuD84vJvURs2ey+1obPvA8jMRyLiPwM/KK77q8yc2WkuSYtaq9pLL+dyuK2qJAlwW1VJ0gKZLCRJpUwWkqRSJgtJUimThSSplMlCklTKZCFJKmWykCSVMllIkkq5raoktdDO/truwS1JS1g7+2svpT24bYaSpCZm2197LtcsFiYLSWqinf21e7EHd78wWUhSE+3sr92LPbj7hclCkppoZ3/tXuzB3S/s4JakJuod1LONdGrnmsXCzY8kSYCbH0mSFshmKElaACflSZJmtZQm5ZksJPWdQfm1PtukvH6MdyFMFpL6yiD9WndSniRVZJCW0HBSniRVZJB+rS+lSXkmC0l9pZ9+rW/eOc7qjVs5ccM3Wb1xK5t3jk87v3bVKG87bZRlEQAsi+Btp432XXNZJ5gsJPWVfvm1Xu87GZ+YJHm276QxYWzeOc7Xd4yzv5jcvD+Tr+8YPyipLAYmC0l9Ze2qUS477xRGR4YJYHRkmMvOO6Xnv9Zdony6SkZDRcSHgH8PJLAbeB9wDPBV4MXADuA9mflURDwX+BJwGvBPwDsy874q4pbUG2tXVd+U4xLl0/W8ZhERo8CfA2OZeTKwDHgn8Eng8sx8GfAosK54yTrg0aL88uI6SeoqlyifrqpmqEOB4Yg4FDgMeBA4A7ixOH8tsLZ4fm5xTHH+zIiiN0mSusQlyqfreTNUZo5HxF8DvwAmgW9Ta3aayMyni8v2AvU66CjwQPHapyPiMWpNVb9ufN+IuBC4EOCEE07o9n+GpEXOJcqn63myiIgjqdUWTgQmgBuAsxf6vpl5FXAV1JYoX+j7SVI7fSf90L/SC1U0Q/074OeZuS8zp4CbgNXASNEsBXAcUB97Ng4cD1CcfyG1jm5JUo9UkSx+AZweEYcVfQ9nAncD24C3F9dcANxcPL+lOKY4vzUX445NktTHep4sMvMOah3Vd1IbNnsIteaji4EPR8S91Pokrilecg3w4qL8w8CGXscsSUud26pKkgC3VZUkLZDJQpJUymQhSSplspAklTJZSJJKmSwkSaVMFpKkUiYLSVIpk4UkqZTJQpJUqpJtVSVpoTbvHJ91H4my85obk4WkgbN55ziX3LSbyan9AIxPTHLJTbuB2v4Sm3eOs/6Gu5h6Jg+cX3/DXQfOa+5cSFDSwFm9cSvjE5MHlS+L4JniO63ZN9vI8BC7Lj2ry9ENrtkWErRmIWng/LJJogDYX/Ljd2JyqhvhLAl2cEsaOMeODFcdwpJjspA0cNavWcnw0LI5v+7Iw4a6EM3SYDOUpIFT76Suj3Y6JKK0CWpoWXDp772yF+EtSiYLSQNp7arRA0lj5ugogKFDgsOfdygTT0w5dLYDTBaSBt7MmobJofNMFpIWhcaaRqc4se9ZJgtJaqJs4t9S42goSWpi05Y90/pAACan9rNpy56KIqqWyUKSmmg18a9V+WJnspCkJlpN/FuqEwJNFpLURLOJf8NDy1i/ZmVFEVWrrWQRERe1UyZJi8XaVaNcdt4pjI4ME8DoyDCXnXfKkuzchjZXnY2IOzPz1TPKdmbmqq5FtgCuOitJczfvVWcj4g+BPwJOjIhbGk4dATyygIBGgKuBk6mtJPwnwB7gemAFcB9wfmY+GhEBXAG8GXgCeG9m3jnfz5YkzV3ZPIv/AzwIHAV8uqH8ceCHC/jcK4B/yMy3R8RzgMOAjwC3ZebGiNgAbAAuBt4EnFQ8XgdcWfyVJPXIrMkiM+8H7gd+u1MfGBEvBH4XeG/xGU8BT0XEucDri8uuBW6nlizOBb6Utfay70fESEQck5kPdiomSdLs2u3gPi8ifhoRj0XEbyLi8Yj4zTw/80RgH/CFiNgZEVdHxPOBoxsSwK+Ao4vno8ADDa/fW5RJknqk3aGznwLOycwXZuYLMvOIzHzBPD/zUODVwJVFB/m/UGtyOqCoRcxpv9eIuDAitkfE9n379s0zNElSM+0mi4cy854OfeZeYG9m3lEc30gteTwUEccAFH8fLs6PA8c3vP64omyazLwqM8cyc2z58uUdClWSBOWjoc4rnm6PiOuBzcCT9fOZedNcPzAzfxURD0TEyszcA5wJ3F08LgA2Fn9vLl5yC/CBiPgqtY7tx+yvkKTeKhsN9XsNz58Azmo4TmDOyaLwH4DripFQPwPeR62W87WIWEetU/384tpvURs2e28Rw/vm+ZmSpHkqGw3VlS/mzNwFNJv4cWaTaxN4fzfikCS1p639LCLib5oUPwZsz8ybm5yTpLa5yVD/a7eD+3nAqcBPi8dvUetoXhcRn+1SbJKWgPomQ+MTkyTPbjK0eedB41hUoXZ3yvstYHVm7geIiCuB/wX8DrC7S7FJWgJm22TI2kX/aLdmcSRweMPx84EXFcnjyeYvkaRybjI0GNqtWXwK2BURtwNBbbmO/1rMvP6fXYpN0hJw7Mgw400Sw1LdZKhftVWzyMxrgH9DbZ7F/wB+JzOvzsx/ycz13QxQ0uLmJkODoWxS3ssz8ycRUd/Lor5G00si4iUuFS5poer9Eo6G6m9lzVAfBi5k+vLkdQmc0fGIJA2khQx/Xbtq1OTQ58om5V1Y/H1Db8KRNIjqw1/ro5rqw18Bk8Ai0e4S5YdFxEcj4qri+KSIeGt3Q5M0KGYb/qrFod2hs18AnqLWyQ21VV8/0ZWIJA0ch78ufu0mi5dm5qeAKYDMfILaEFpJajnM1eGvi0e7yeKpiBim2JAoIl6Kk/EkFRz+uvi1OynvUuAfgOMj4jpgNcUe2pLk8NfFL2orgJdcFPFl4IfAJLX9J+7IzF93ObZ5Gxsby+3bt1cdhiQNlIjYkZnNto9ou2ZxDfBvgTcCLwV2RsR3MvOKDsUoSepjbSWLzNwWEd8BXgO8Afgz4JWAyUKSloB2Nz+6jdpKs9+jtjT5azLz4W4GJknqH+2OhvohtXkWJ1Pb2+LkYnSUJGkJaLcZ6kMAEXEEtVFQXwBeAjy3a5FJkvpGu81QH6DWwX0acB/weWrNUZKkJaDd0VDPAz4D7MjMp7sYjySpD7XbDPXX3Q5EktS/2u3gliQtYSYLSVIpk4UkqZTJQpJUymQhSSpVWbKIiGURsTMivlEcnxgRd0TEvRFxfUQ8pyh/bnF8b3F+RVUxS9JSVWXN4iLgnobjTwKXZ+bLgEeBdUX5OuDRovzy4jpJUg9Vkiwi4jjgLcDVxXEAZwA3FpdcC6wtnp9bHFOcP7O4XpLUI1XVLD4L/CXwTHH8YmCiYXb4XqC+xdYo8ABAcf6x4vppIuLCiNgeEdv37dvXzdglacnpebKIiLcCD2fmjk6+b2ZelZljmTm2fPnyTr61JC157a4N1UmrgXMi4s3U1px6AbVNlEYi4tCi9nAcMF5cPw4cD+yNiEOBFwL/1PuwJWnp6nnNIjMvyczjMnMF8E5ga2a+C9gGvL247ALg5uL5LcUxxfmt2c7G4ZKkjumneRYXAx+OiHup9UlcU5RfA7y4KP8wsKGi+CRpyaqiGeqAzLwduL14/jPgtU2u+X/AH/Q0MEnSNP1Us5Ak9SmThSSplMlCklTKZCFJKmWykCSVqnQ0lKTe2bxznE1b9vDLiUmOHRlm/ZqVrF01Wv5CCZOFtCRs3jnOJTftZnJqPwDjE5NcctNuABOG2mIzlLQEbNqy50CiqJuc2s+mLXsqikiDxmQhLQG/nJicU7k0k81QUh+Za79Cu9cfOzLMeJPEcOzIcEfj1+JlzULqE/V+hfGJSZJn+xU27xxf8PXr16xkeGjZtLLhoWWsX7OyC/8lWoxMFlKfmGu/wlyuX7tqlMvOO4XRkWECGB0Z5rLzTrFzW22zGUrqE3PtV5hr+dpVoyYHzZs1C6lPtOo/6FS5tBAmC6lPzLVfwX4I9ZLNUFKfqDcRbdqyh/GJSZZFTOuDmNmE1Hi9s7LVbSYLqY/Uv+jX33gXU/truwePT0yy/sa7pp1vvN7koF6wGUrqMx//+x8fSBR1U/uTj//9jyuKSDJZSH3n0SemWpafuOGbrN64teXcC6lbTBbSAGlnsp7UDSYLqc9ElF/jIoDqNZOF1Gcyy68BFwFUbzkaSppFFRsGjbZY9G8mJ9+pl6xZSC3MdWG/Tmk22W4mJ9+p10wWUgvd2jBo885xVm/c2nJkU7NF/959+gkuAqhK2QwltdCNDYPa3d7UyXbqN9YspBa6sVCf25tqUPU8WUTE8RGxLSLujogfR8RFRfmLIuLWiPhp8ffIojwi4m8i4t6I+GFEvLrXMWtpWshCfa2amtzeVIOqimaop4G/yMw7I+IIYEdE3Aq8F7gtMzdGxAZgA3Ax8CbgpOLxOuDK4q/UVfNdqG+2pia3N9Wg6nmyyMwHgQeL549HxD3AKHAu8PrismuB26kli3OBL2VmAt+PiJGIOKZ4H6mrmvUdlA2nna2paf2aldMSCTiySYOh0g7uiFgBrALuAI5uSAC/Ao4uno8CDzS8bG9RNi1ZRMSFwIUAJ5xwQtdi1tLWrNbwoet38cHrdzFaJI7ZmppcVlyDqrJkERGHA18HPpiZv4mGNQ4yMyOizXmsB15zFXAVwNjY2JxeK7WrWa2h/j9bvblp5LChposB1puaHOmkQVRJsoiIIWqJ4rrMvKkofqjevBQRxwAPF+XjwPENLz+uKJM6rqyJqawjenJqP8899BCGh5bZ1KRFpYrRUAFcA9yTmZ9pOHULcEHx/ALg5obyPy5GRZ0OPGZ/hbqhnRnb7XREPzY5ddCkOifRadBVUbNYDbwH2B0Ru4qyjwAbga9FxDrgfuD84ty3gDcD9wJPAO/rbbhaKmbrmD6wg12TDuqZjh0ZtqlJi04Vo6H+N9BqEeYzm1yfwPu7GpREe3MgZu6THTzbZwE2N2nxcrkPLRkz+yPe8PLlbPvJvgPHZR3TdY21hipWpZWqENnu4vkDZGxsLLdv3151GOojM4e8NjN0SEAwbf/roUOCw593KBNPTJkMtOhFxI7MHGt2zrWhtCQ064+YaeqZ5PnPOfRAx/TI8BBEbe9rtzPVUmczlAZeq6agj27ezVfueID9c6g9PzY5xa5LzwJg9catTExOb5aa2eEtLRUmCw2keoKY2clc//V/w/Zf8N1/fGTO79vYP+Gif9KzbIbSwGmcDwHTRyNB7df/fBLFzJFM3ViiXBpUJgsNnHb6H+bqyMOGDpo4t5AlyqXFxmYoDZxmS3zP17IIPn3+q5r2Qbjon/Qsk4UGziEBz8zSZz1zXabZritbhsOZ2FKNzVDquFa7xHXKbImivg7TbFyvSZo7axbqqNl2ievFF/N3N5wBwAev39Xymp9vfEvX45AWG2sW6qjZFuPrlJHhodLy0RYjllqVS5qdNQt11HznJpSt29TYsfyxc17J+hvuYuqZ6ctyfOycVx44dvtSqbNMFuqoY0eGm45WOnZkeNpEumUR7M9ktEgMX98xPq3p6svf/8WB185sympnlJIjmaTOciFBzVuzZTaApr/o33ba6LSE0GjmMt+tjI4MH+iTkNR5LiSojmu1qxzQdJe4bT/Z13I4a7s/V1xmQ6qOzVCal9k6sr+74YyDmns+NMvopHa5zIZUHWsWmpe5dmSXfdG32jqxzs5pqVomCx2knUl1c11kr9k6S3XDQ8t41+knTGu6eveMYyfQSdWyGUrTtDupbq5DU2fuXd04GspRSlL/M1ksIe3sFz1bX8RCh6a6zpI0uEwWA66dBFC/rp0aw1z6Ivzyl5YO+ywGWKvhq836GNpdhsMNfyQ1Y7IYYHNZh6ndGoMb/khqxmQxwObSZNRujWHtqtGmk+psbpKWNvssBths6zDNNJfRS/ZFSJrJmkWDbm/a02lzaTKyxiBpIQamZhERZwNXAMuAqzNzYyffv+pNe+ZjrsNXrTFImq+BSBYRsQz4W+CNwF7gBxFxS2be3anPaHd+Qb8xAUjqhUFphnotcG9m/iwznwK+CpzbyQ+Y76Y9krQUDEqyGAUeaDjeW5QdEBEXRsT2iNi+b9++OX+A8wskqbVBSRalMvOqzBzLzLHly5fP+fXOL5Ck1gaizwIYB45vOD6uKOsYt+GUpNYGJVn8ADgpIk6kliTeCfxRpz/EzmJJam4gkkVmPh0RHwC2UBs6+/nM/HHFYUnSkjEQyQIgM78FfKvqOCRpKVo0HdySpO4xWUiSSpksJEmlIjOrjqHjImIfcH+PP/Yo4Nc9/sz5MtbOG5Q4wVi7ZVBinS3Of52ZTSeqLcpkUYWI2J6ZY1XH0Q5j7bxBiROMtVsGJdb5xmkzlCSplMlCklTKZNE5V1UdwBwYa+cNSpxgrN0yKLHOK077LCRJpaxZSJJKmSwkSaVMFvMQEZ+PiIcj4kcNZS+KiFsj4qfF3yOrjLGuRawfi4jxiNhVPN5cZYxFTMdHxLaIuDsifhwRFxXlfXdfZ4m1H+/r8yLi/0bEXUWsHy/KT4yIOyLi3oi4PiKe06dxfjEift5wT0+tMs5GEbEsInZGxDeK4766p42axDrn+2qymJ8vAmfPKNsA3JaZJwG3Fcf94IscHCvA5Zl5avHohwUanwb+IjNfAZwOvD8iXkF/3tdWsUL/3dcngTMy81XAqcDZEXE68Elqsb4MeBRYV2GM0DpOgPUN93RXdSEe5CLgnobjfrunjWbGCnO8ryaLecjM7wCPzCg+F7i2eH4tsLanQbXQIta+k5kPZuadxfPHqf2PPUof3tdZYu07WfPPxeFQ8UjgDODGorzy+zpLnH0pIo4D3gJcXRwHfXZP62bGOl8mi845OjMfLJ7/Cji6ymDa8IGI+GHRTFV5006jiFgBrALuoM/v64xYoQ/va9EEsQt4GLgV+EdgIjOfLi45aE/7KsyMMzPr9/S/FPf08oh4boUhNvos8JfAM8Xxi+nDe1qYGWvdnO6ryaILsjYeuW9/FQFXAi+lVt1/EPh0teE8KyIOB74OfDAzf9N4rt/ua5NY+/K+Zub+zDyV2nbErwVeXnFITc2MMyJOBi6hFu9rgBcBF1cYIgAR8Vbg4czcUXUsZWaJdc731WTROQ9FxDEAxd+HK46npcx8qPiH+QzwOWpfIJWLiCFqX77XZeZNRXFf3tdmsfbrfa3LzAlgG/DbwEhE1Dc/6/ie9gvREOfZRZNfZuaTwBfoj3u6GjgnIu4Dvkqt+ekK+vOeHhRrRHx5PvfVZNE5twAXFM8vAG6uMJZZ1b98C78P/KjVtb1StPleA9yTmZ9pONV397VVrH16X5dHxEjxfBh4I7U+lm3A24vLKr+vLeL8ScMPhaDWB1D5Pc3MSzLzuMxcAbwT2JqZ76LP7im0jPXd87mvA7Otaj+JiK8ArweOioi9wKXARuBrEbGO2vLo51cX4bNaxPr6YqhcAvcBf1pZgM9aDbwH2F20WwN8hP68r61i/cM+vK/HANdGxDJqPw6/lpnfiIi7ga9GxCeAndSSX5Vaxbk1IpYDAewC/qzKIEtcTH/d09lcN9f76nIfkqRSNkNJkkqZLCRJpUwWkqRSJgtJUimThSSplMlCmoeI+Ofyq+b8nqc2rlRbrGL7Hzv9OdJ8mCyk/nEqUPmy5lIzJgtpgSJifUT8oFiUrb4Pw4qIuCciPlfsz/DtYmYyEfGa4tpdEbEpIn5U7H3wV8A7ivJ3FG//ioi4PSJ+FhF/XtF/omSykBYiIs4CTqK2ts6pwGkR8bvF6ZOAv83MVwITwNuK8i8Af1osmrcfIDOfAv4TcH2xv8D1xbUvB9YU739psSaV1HMmC2lhzioeO4E7qX25n1Sc+3nDpjI7gBXF+kdHZOb3ivK/K3n/b2bmk5n5a2qLKPbVEu1aOlwbSlqYAC7LzP8+rbC2z8WTDUX7geF5vP/M9/DfrCphzUJamC3AnxR7WxARoxHxr1pdXCy//XhEvK4oemfD6ceBI7oWqbQAJgtpATLz29Sakr4XEbupbatZ9oW/DvhcsWLt84HHivJt1Dq0Gzu4pb7gqrNSj0XE4fX9piNiA3BMZl5UcVjSrGz/lHrvLRFxCbV/f/cD7602HKmcNQtJUin7LCRJpUwWkqRSJgtJUimThSSplMlCklTq/wPl5yBP2W4mzwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "(42,) (14,)\n",
            "(42, 1) (14, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WVD8SVTeHOLu"
      },
      "source": [
        "### 결정계수 $R^2$\n",
        "\n",
        "KNeighborsRegressor: k-최근접 이웃 회귀 알고리즘을 구현한 클래스"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eJ3z5pkWIFxU"
      },
      "source": [
        "$ R^2 = 1 - \\frac{\\sum(타깃 - 예측)^{2}}{\\sum(타깃 - 평균)^{2}}$"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7HdeRynzHWTV",
        "outputId": "3cc39a9d-cd64-4f4b-e96e-01969875c1a1"
      },
      "source": [
        "from sklearn.neighbors import KNeighborsRegressor\n",
        "\n",
        "# k-최근접 이웃 회귀 알고리즘\n",
        "knr = KNeighborsRegressor()\n",
        "# 훈련\n",
        "knr.fit(train_input, train_target)\n",
        "# 평가\n",
        "print(knr.score(test_input, test_target)) # 결정 계수 R^2"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9928094061010639\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "U6vCh2DwJ1Md",
        "outputId": "0beb7c87-29fd-4152-ce1c-5e3a4064af01"
      },
      "source": [
        "from sklearn.metrics import mean_absolute_error\n",
        "\n",
        "# 예측\n",
        "test_prediction = knr.predict(test_input)\n",
        "# 평균 절댓값 오차를 계산\n",
        "mae = mean_absolute_error(test_target, test_prediction)\n",
        "print(mae)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "19.157142857142862\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IKDulrMUKfEu"
      },
      "source": [
        "### 과대적합 vs 과소적합"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EA26An65KoGc",
        "outputId": "82b77e9e-800e-4ef1-8c5f-c32cea640e5c"
      },
      "source": [
        "print(knr.score(train_input, train_target)) \n",
        "# 훈련세트 < 테스트세트 -> 과소적합\n",
        "\n",
        "# k-최근접 이웃 알고리즘으로, 모델 복잡하게 만드는 방법: 이웃의 개수 k를 줄이기\n",
        "\n",
        "# 이웃의 갯수를 3으로 설정\n",
        "knr.n_neighbors = 3\n",
        "# 모델을 다시 훈련\n",
        "knr.fit(train_input, train_target)\n",
        "print(knr.score(train_input, train_target))\n",
        "print(knr.score(test_input, test_target))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9698823289099255\n",
            "0.9804899950518966\n",
            "0.974645996398761\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BpVZN12kL--X"
      },
      "source": [
        "## 03-2. 선형 회귀"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CosW29EKMCn6"
      },
      "source": [
        "### k-최근접 이웃의 한계\n",
        "농어가 아무리 커도 무게가 더 늘어나지 않는다!!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HhWInkOKqg6s",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 575
        },
        "outputId": "63f0b8ce-8183-40b0-b7b4-5d3b5ee60093"
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "perch_length = np.array(\n",
        "    [8.4, 13.7, 15.0, 16.2, 17.4, 18.0, 18.7, 19.0, 19.6, 20.0, \n",
        "     21.0, 21.0, 21.0, 21.3, 22.0, 22.0, 22.0, 22.0, 22.0, 22.5, \n",
        "     22.5, 22.7, 23.0, 23.5, 24.0, 24.0, 24.6, 25.0, 25.6, 26.5, \n",
        "     27.3, 27.5, 27.5, 27.5, 28.0, 28.7, 30.0, 32.8, 34.5, 35.0, \n",
        "     36.5, 36.0, 37.0, 37.0, 39.0, 39.0, 39.0, 40.0, 40.0, 40.0, \n",
        "     40.0, 42.0, 43.0, 43.0, 43.5, 44.0]\n",
        "     )\n",
        "perch_weight = np.array(\n",
        "    [5.9, 32.0, 40.0, 51.5, 70.0, 100.0, 78.0, 80.0, 85.0, 85.0, \n",
        "     110.0, 115.0, 125.0, 130.0, 120.0, 120.0, 130.0, 135.0, 110.0, \n",
        "     130.0, 150.0, 145.0, 150.0, 170.0, 225.0, 145.0, 188.0, 180.0, \n",
        "     197.0, 218.0, 300.0, 260.0, 265.0, 250.0, 250.0, 300.0, 320.0, \n",
        "     514.0, 556.0, 840.0, 685.0, 700.0, 700.0, 690.0, 900.0, 650.0, \n",
        "     820.0, 850.0, 900.0, 1015.0, 820.0, 1100.0, 1000.0, 1100.0, \n",
        "     1000.0, 1000.0]\n",
        "     )\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "train_input, test_input, train_target, test_target = train_test_split(perch_length, perch_weight, random_state=42)\n",
        "train_input = train_input.reshape(-1, 1)\n",
        "test_input = test_input.reshape(-1, 1)\n",
        "\n",
        "from sklearn.neighbors import KNeighborsRegressor\n",
        "\n",
        "knr = KNeighborsRegressor(n_neighbors=3)\n",
        "knr.fit(train_input, train_target)\n",
        "\n",
        "print(knr.predict([[50]]))\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# 50cm 농어의 이웃을 구합니다\n",
        "distances, indexes = knr.kneighbors([[50]])\n",
        "\n",
        "# 훈련 세트의 산점도를 그립니다\n",
        "plt.scatter(train_input, train_target)\n",
        "# 훈련 세트 중에서 이웃 샘플만 다시 그립니다\n",
        "plt.scatter(train_input[indexes], train_target[indexes], marker='D')\n",
        "# 50cm 농어 데이터\n",
        "plt.scatter(50, 1033, marker='^')\n",
        "plt.xlabel('length')\n",
        "plt.ylabel('weight')\n",
        "plt.show()\n",
        "\n",
        "print(np.mean(train_target[indexes]))\n",
        "\n",
        "# 100cm 농어의 이웃을 구합니다\n",
        "distances, indexes = knr.kneighbors([[100]])\n",
        "\n",
        "# 훈련 세트의 산점도를 그립니다\u001f\n",
        "plt.scatter(train_input, train_target)\n",
        "# 훈련 세트 중에서 이웃 샘플만 다시 그립니다\u001f\n",
        "plt.scatter(train_input[indexes], train_target[indexes], marker='D')\n",
        "# 100cm 농어 데이터\n",
        "plt.scatter(100, 1033, marker='^')\n",
        "plt.xlabel('length')\n",
        "plt.ylabel('weight')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[1033.33333333]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "1033.3333333333333\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GMx5ZCCbNnEy"
      },
      "source": [
        "### 선형 회귀"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 347
        },
        "id": "4jase26WN1N4",
        "outputId": "15729aea-48c2-4ca7-c127-1d33d17af40e"
      },
      "source": [
        "from sklearn.linear_model import LinearRegression\n",
        "\n",
        "lr = LinearRegression()\n",
        "# 선형 회귀 모델 훈련\n",
        "lr.fit(train_input, train_target)\n",
        "\n",
        "# 50cm 농어에 대한 예측\n",
        "print(lr.predict([[50]]))\n",
        "\n",
        "print(lr.coef_, lr.intercept_)\n",
        "\n",
        "# 훈련 세트의 산점도를 그립니다\u001f\n",
        "plt.scatter(train_input, train_target)\n",
        "# 15에서 50까지 1차 방정식 그래프를 그립니다\n",
        "plt.plot([15, 50], [15*lr.coef_+lr.intercept_, 50*lr.coef_+lr.intercept_])\n",
        "# 50cm 농어 데이터\n",
        "plt.scatter(50, 1241.8, marker='^')\n",
        "plt.xlabel('length')\n",
        "plt.ylabel('weight')\n",
        "plt.show()\n",
        "\n",
        "print(lr.score(train_input, train_target))\n",
        "print(lr.score(test_input, test_target))\n",
        "# 훈련 세트 > 테스트 세트 이지만, 과대적합 x\n",
        "# 훈련 세트도 낮음 -> 과소적합 o"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[1241.83860323]\n",
            "[39.01714496] -709.0186449535477\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "0.9398463339976039\n",
            "0.8247503123313558\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kLPMSXUpOgAG"
      },
      "source": [
        "### 다항 회귀"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 330
        },
        "id": "w7y-r3WeOk2B",
        "outputId": "92922304-3537-4a12-ea22-77be2e5da74b"
      },
      "source": [
        "train_poly = np.column_stack((train_input ** 2, train_input))\n",
        "test_poly = np.column_stack((test_input ** 2, test_input))\n",
        "\n",
        "print(train_poly.shape, test_poly.shape)\n",
        "\n",
        "lr = LinearRegression()\n",
        "lr.fit(train_poly, train_target)\n",
        "\n",
        "print(lr.predict([[50**2, 50]]))\n",
        "\n",
        "print(lr.coef_, lr.intercept_)\n",
        "\n",
        "# 구간별 직선을 그리기 위해 15에서 49까지 정수 배열을 만듭니다\n",
        "point = np.arange(15, 50)\n",
        "# 훈련 세트의 산점도를 그립니다\n",
        "plt.scatter(train_input, train_target)\n",
        "# 15에서 49까지 2차 방정식 그래프를 그립니다\n",
        "plt.plot(point, 1.01*point**2 - 21.6*point + 116.05)\n",
        "# 50cm 농어 데이터\n",
        "plt.scatter([50], [1574], marker='^')\n",
        "plt.xlabel('length')\n",
        "plt.ylabel('weight')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(42, 2) (14, 2)\n",
            "[1573.98423528]\n",
            "[  1.01433211 -21.55792498] 116.05021078278276\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "i7DCD7XpO9z4",
        "outputId": "fb41555f-7ab3-4d80-95db-5ea20c37ae0a"
      },
      "source": [
        "print(lr.score(train_poly, train_target))\n",
        "print(lr.score(test_poly, test_target))\n",
        "# 훈련 세트 < 테스트 세트 -> 과소적합"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9706807451768623\n",
            "0.9775935108325122\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KJydoAdKPJ1m"
      },
      "source": [
        "## 03-3. 특성 공학과 규제"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DySYPGTgPZaW"
      },
      "source": [
        "### 다중 회귀\n",
        "- 특성 공학: 기존의 특성을 사용해 새로운 특성을 뽑아내는 작업\n",
        "- 판다스: 데이터 분석 라이브러리\n",
        "- 데이터프레임: 판다스의 핵심 데이터 구조\n",
        "  - 다차원 배열 다룰 수 있음\n",
        "  - 데이터프레임을 넘파이 배열로 쉽게 바꿀 수 있음\n",
        "  - CSV 파일: 판다스 데이터프레임을 만들기 위해 많이 사용하는 파일\n",
        "\n",
        "> pd.read_csv(): CSV 파일 > 판다스 데이터프레임  \n",
        "> to_numpy(): 판다스 데이터 프레임 > 넘파이 배열"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "APissJksPYNo",
        "outputId": "51458e8b-8a32-4cb5-ea91-30681857217a"
      },
      "source": [
        "import pandas as pd\n",
        "\n",
        "df = pd.read_csv('https://bit.ly/perch_csv_data')\n",
        "perch_full = df.to_numpy()\n",
        "print(perch_full)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[ 8.4   2.11  1.41]\n",
            " [13.7   3.53  2.  ]\n",
            " [15.    3.82  2.43]\n",
            " [16.2   4.59  2.63]\n",
            " [17.4   4.59  2.94]\n",
            " [18.    5.22  3.32]\n",
            " [18.7   5.2   3.12]\n",
            " [19.    5.64  3.05]\n",
            " [19.6   5.14  3.04]\n",
            " [20.    5.08  2.77]\n",
            " [21.    5.69  3.56]\n",
            " [21.    5.92  3.31]\n",
            " [21.    5.69  3.67]\n",
            " [21.3   6.38  3.53]\n",
            " [22.    6.11  3.41]\n",
            " [22.    5.64  3.52]\n",
            " [22.    6.11  3.52]\n",
            " [22.    5.88  3.52]\n",
            " [22.    5.52  4.  ]\n",
            " [22.5   5.86  3.62]\n",
            " [22.5   6.79  3.62]\n",
            " [22.7   5.95  3.63]\n",
            " [23.    5.22  3.63]\n",
            " [23.5   6.28  3.72]\n",
            " [24.    7.29  3.72]\n",
            " [24.    6.38  3.82]\n",
            " [24.6   6.73  4.17]\n",
            " [25.    6.44  3.68]\n",
            " [25.6   6.56  4.24]\n",
            " [26.5   7.17  4.14]\n",
            " [27.3   8.32  5.14]\n",
            " [27.5   7.17  4.34]\n",
            " [27.5   7.05  4.34]\n",
            " [27.5   7.28  4.57]\n",
            " [28.    7.82  4.2 ]\n",
            " [28.7   7.59  4.64]\n",
            " [30.    7.62  4.77]\n",
            " [32.8  10.03  6.02]\n",
            " [34.5  10.26  6.39]\n",
            " [35.   11.49  7.8 ]\n",
            " [36.5  10.88  6.86]\n",
            " [36.   10.61  6.74]\n",
            " [37.   10.84  6.26]\n",
            " [37.   10.57  6.37]\n",
            " [39.   11.14  7.49]\n",
            " [39.   11.14  6.  ]\n",
            " [39.   12.43  7.35]\n",
            " [40.   11.93  7.11]\n",
            " [40.   11.73  7.22]\n",
            " [40.   12.38  7.46]\n",
            " [40.   11.14  6.63]\n",
            " [42.   12.8   6.87]\n",
            " [43.   11.93  7.28]\n",
            " [43.   12.51  7.42]\n",
            " [43.5  12.6   8.14]\n",
            " [44.   12.49  7.6 ]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1t-9HPhMRneY"
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "perch_weight = np.array(\n",
        "    [5.9, 32.0, 40.0, 51.5, 70.0, 100.0, 78.0, 80.0, 85.0, 85.0, \n",
        "     110.0, 115.0, 125.0, 130.0, 120.0, 120.0, 130.0, 135.0, 110.0, \n",
        "     130.0, 150.0, 145.0, 150.0, 170.0, 225.0, 145.0, 188.0, 180.0, \n",
        "     197.0, 218.0, 300.0, 260.0, 265.0, 250.0, 250.0, 300.0, 320.0, \n",
        "     514.0, 556.0, 840.0, 685.0, 700.0, 700.0, 690.0, 900.0, 650.0, \n",
        "     820.0, 850.0, 900.0, 1015.0, 820.0, 1100.0, 1000.0, 1100.0, \n",
        "     1000.0, 1000.0]\n",
        "     )\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "train_input, test_input, train_target, test_target = train_test_split(perch_full, perch_weight, random_state=42)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E_W-A-RmRWsW"
      },
      "source": [
        "### 사이킷런의 변환기"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eD7HohPGRaM7"
      },
      "source": [
        "변환기: 특성 만들거나 전처리하는 클래스\n",
        "PolynomialFeatures 클래스"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_xLud9ChRZzz",
        "outputId": "4151da09-7a6e-42d8-d0cb-e1351eeeab28"
      },
      "source": [
        "from sklearn.preprocessing import PolynomialFeatures\n",
        "\n",
        "# 특성 2, 3으로 특성 만들기\n",
        "poly = PolynomialFeatures()\n",
        "poly.fit([[2, 3]])\n",
        "print(poly.transform([[2, 3]]))\n",
        "\n",
        "# 선형 방정식의 절편 값이 1인 특성과 곱해지는 계수\n",
        "# 사이킷런의 선형 모델은 자동으로 절편을 추가함 -> 1인 특성을 만들 필요 없음\n",
        "poly = PolynomialFeatures(include_bias=False)\n",
        "poly.fit([[2, 3]])\n",
        "print(poly.transform([[2, 3]]))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[[1. 2. 3. 4. 6. 9.]]\n",
            "[[2. 3. 4. 6. 9.]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wsGeH7oaSvkZ",
        "outputId": "06a607b1-0003-4e5c-a595-2d4f83e1ffa4"
      },
      "source": [
        "poly = PolynomialFeatures(include_bias=False)\n",
        "\n",
        "poly.fit(train_input)\n",
        "train_poly = poly.transform(train_input)\n",
        "\n",
        "print(train_poly.shape)\n",
        "print(poly.get_feature_names())\n",
        "\n",
        "# 훈련 세트와 함께 테스트 세트 변환\n",
        "test_poly = poly.transform(test_input)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(42, 9)\n",
            "['x0', 'x1', 'x2', 'x0^2', 'x0 x1', 'x0 x2', 'x1^2', 'x1 x2', 'x2^2']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-OUmZSJcTC4J"
      },
      "source": [
        "### 다중 회귀 모델 훈련하기 -> 여기서부터 다시 보기"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AKeP-ZOhVV_B"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lw6S51VzVWS_"
      },
      "source": [
        "import numpy as np\n",
        "\n",
        "perch_weight = np.array(\n",
        "    [5.9, 32.0, 40.0, 51.5, 70.0, 100.0, 78.0, 80.0, 85.0, 85.0, \n",
        "     110.0, 115.0, 125.0, 130.0, 120.0, 120.0, 130.0, 135.0, 110.0, \n",
        "     130.0, 150.0, 145.0, 150.0, 170.0, 225.0, 145.0, 188.0, 180.0, \n",
        "     197.0, 218.0, 300.0, 260.0, 265.0, 250.0, 250.0, 300.0, 320.0, \n",
        "     514.0, 556.0, 840.0, 685.0, 700.0, 700.0, 690.0, 900.0, 650.0, \n",
        "     820.0, 850.0, 900.0, 1015.0, 820.0, 1100.0, 1000.0, 1100.0, \n",
        "     1000.0, 1000.0]\n",
        "     )\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "train_input, test_input, train_target, test_target = train_test_split(perch_full, perch_weight, random_state=42)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ct0rGeclTF78",
        "outputId": "e4a5f446-885d-4f91-c50c-b913f2a94f7f"
      },
      "source": [
        "from sklearn.linear_model import LinearRegression\n",
        "\n",
        "lr = LinearRegression()\n",
        "lr.fit(train_poly, train_target)\n",
        "print(lr.score(train_poly, train_target))\n",
        "print(lr.score(test_poly, test_target))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9903183436982124\n",
            "0.9714559911594132\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nGLdb7_zTU9i",
        "outputId": "2288daf8-5ead-46fb-e274-fc29d1bedbf8"
      },
      "source": [
        "poly = PolynomialFeatures(degree=5, include_bias=False)\n",
        "\n",
        "poly.fit(train_input)\n",
        "train_poly = poly.transform(train_input)\n",
        "test_poly = poly.transform(test_input)\n",
        "\n",
        "print(train_poly.shape)\n",
        "\n",
        "lr.fit(train_poly, train_target)\n",
        "print(lr.score(train_poly, train_target))\n",
        "\n",
        "print(lr.score(test_poly, test_target)) # 음수 -> 과대적합"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(42, 55)\n",
            "0.9999999999991096\n",
            "-144.40579242335605\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HbKcjq8ET2Ju"
      },
      "source": [
        "### 규제\n",
        "규제: 과대적합 감소, 선형 회귀 모델의 경우 큭성에 곱해지는 계수(or 기울기)의 크기를 작게 만드는 일\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yOb3BLfpUWC5"
      },
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "ss = StandardScaler()\n",
        "ss.fit(train_poly)\n",
        "\n",
        "train_scaled = ss.transform(train_poly)\n",
        "test_scaled = ss.transform(test_poly)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cXrXuryvT7Uk"
      },
      "source": [
        "### 릿지 회귀\n",
        "과대적합 감소  \n",
        "alpha 감소 -> 규제 증가 -> 계수 감소"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bcRYXkyxVp0j"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MzsrhabmUZsI",
        "outputId": "51d84247-24d5-4e22-e747-f2914dae500a"
      },
      "source": [
        "from sklearn.linear_model import Ridge\n",
        "\n",
        "ridge = Ridge()\n",
        "ridge.fit(train_scaled, train_target)\n",
        "print(ridge.score(train_scaled, train_target))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9896101671037343\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uS5iuhyAUbm8",
        "outputId": "527e5eea-315d-4984-8e34-7aeea2e5eaba"
      },
      "source": [
        "print(ridge.score(test_scaled, test_target))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9790693977615398\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LBejJLgjUdA5"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "train_score = []\n",
        "test_score = []"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kuFVdlQ8Ueoa"
      },
      "source": [
        "alpha_list = [0.001, 0.01, 0.1, 1, 10, 100]\n",
        "for alpha in alpha_list:\n",
        "    # 릿지 모델을 만듭니다\n",
        "    ridge = Ridge(alpha=alpha)\n",
        "    # 릿지 모델을 훈련합니다\n",
        "    ridge.fit(train_scaled, train_target)\n",
        "    # 훈련 점수와 테스트 점수를 저장합니다\n",
        "    train_score.append(ridge.score(train_scaled, train_target))\n",
        "    test_score.append(ridge.score(test_scaled, test_target))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "fwcIzhMsUgE2",
        "outputId": "800cc56f-d9c2-4e59-f5dd-4fb2c13a5618"
      },
      "source": [
        "plt.plot(np.log10(alpha_list), train_score)\n",
        "plt.plot(np.log10(alpha_list), test_score)\n",
        "plt.xlabel('alpha')\n",
        "plt.ylabel('R^2')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEGCAYAAABy53LJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXxU1fnH8c+TPYGQQBLWsAkooqJo2BRc0WK1CmgtiNatxbZS2/qjC7W/+qutba2obdXaUqBiRVFR3KoiKhZQWYJssoookLBFIGEJIdvz++PcwBACYWAmdzLzvF+veWXm3js3z33Z8s2559xzRFUxxhhjjlec3wUYY4xpXCw4jDHGBMWCwxhjTFAsOIwxxgTFgsMYY0xQEvwuoCFkZ2drp06d/C7DGGMalUWLFn2lqjm1t8dEcHTq1In8/Hy/yzDGmEZFRDbUtd1uVRljjAmKBYcxxpigWHAYY4wJigWHMcaYoFhwGGOMCYoFhzHGmKBYcBhjjAlKTDzHcaJe/qSA7XsO0CYjhbaZqbTJSKFVsxQS4y1vjTGxy4LjGP6zbAvvrd5+2LY4gZz0ZNpkpNI2M4W2Gam0yUylbUbKwZ/ZTZOJixOfqjbGmPCy4DiGibf2Zu+BSrYU72dzSdlhP7eUlLF66x5mrS5if0XVYd9LjBdaNasJFddaaZuRQpuazxmpZKYlImLhYoxpfCw46tE0OYFurdLp1iq9zv2qSsn+CgqL97OluIwtJYeHzCcbd/Hm8i1UVB2+0mJKYtzBYHGtl8NbLW0yU2mabP95jDGRx/5lOkkiQmZaEplpSZzRNqPOY6qrla/2Hqiz1bK5ZD9zP/uK7XvKqK61im96SsJh4dIu8/BWS+uMFFIS4xvgKo0x5hALjgYQFye0bJZCy2YpnNM+s85jKqqq2ba7zIWJFyoHQ6ZkP8sLStixr/yI72U1STrUaqnpxA9otbRKTybBOvONMSFkwREhEuPjyG2eRm7ztKMeU1ZRdXigBATLxh2lzFu/gz1llYd9J06gZXrKwVZK24O3xg61XrKbWGe+Meb4WXA0IimJ8XTObkLn7CZHPWZPWcVRWy2rtuzm3VXbOFBZfdh3kuLjaJWRfLDV0q55Ku29EMtt7vpfkhKs1WKMcSw4okx6SiLpKYmceozO/F2lFYeCpWQ/m71O/S3FZSz8chevL9tCVUCHiwi0bpZCbvPUg2GSGxAubTLt2RZjYokFR4wREVo0SaJFkyTObFd3Z35lVTVbd5dRsGu/9yqlYNd+Nu0sZcEXO3l1yf7DOvLjaoKlRVqd4dImI8X6WYyJIhYc5ggJ9fS3VFRVs7WkjE1eoASGy7zPd7B1d+FhwRIfJwdbLO3rCJfWzSxYjGlMLDhM0BLj42jfIo32LeoOlvJKFywFu0qPCJcP133F1t1laECwJMQJbTJTyM1MqzNcWjVLId46742JGBYcJuSSEuLokJVGh6y6g+VAZRVbissOhklguMz+rIhtuw8cdnxivNA207VO6gqXluk2KsyYhmTBYRpcckI8nbKb0Okoo8PKKqrYXLz/sJbKJu/n+2u2U7Tn8GBJio+jbWZKHbfB0mjfPNXmDjMmxCw4TMRJSYznlJymnJLTtM79ZRVVh/WrBIbLzJXb+Grv4Q9KJiXEkZuZ6oYZ19HHktM02eYNMyYIYQ0OERkM/AWIByao6h9r7e8ITAJygJ3ATapa4O17ELjKO/S3qvq8t70zMBXIAhYBN6vqkY9Um6iVkhhP15ZN6dqy7mDZX15FYbHXStl5eLi8/elWdtZ6Aj85IY7spslkpCaSmeZeGalJ3s9EMlMPbQs8JjUx3gLHxKSwBYeIxANPAJcDBcBCEXlNVVcGHDYOeFpVJ4vIpcAfgJtF5CrgXOAcIBn4QETeUtXdwIPAo6o6VUT+DtwBPBmu6zCNT2pSPF1bptO1Zd3Psuw7UElh8eHDjHfsK2f3/gqKSytYu20vxaUVlOwvP2JyykBJ8XE0qwkS72ez1EQyvdCpCR4XNkkHj0lPSbTOftOohbPF0QdYp6rrAURkKnAtEBgcPYB7vPezgFcCts9W1UqgUkSWAYNF5EXgUuBG77jJwP9hwWGC0CQ5gVNbpR/1Ickaqsr+iiqKS12gFO8/FC7F3s+S/S5giksr2FxcxqoteyjZX8HeA5XHPHezlARvcszAcHGhk5GaSMbBMDrUyslITbRJLU1ECGdwtAM2BXwuAPrWOmYpMAx3O2sokC4iWd72+0TkYSANuAQXOFlAsRcoNedsV9cvF5FRwCiADh06hOJ6TIwREdKSEkhLSqBtZmpQ362oqqakjnA59LmC4tJyir33hbv2e2FUfsQsyYFSEuPqCJdDrZrAkKlp+WSkJZKenGC31UzI+N05PgZ4XERuBWYDhUCVqr4jIr2Bj4Ai4GOg6qhnqYOqjgfGA+Tl5R3j/4rGhF5ivOs3yW6aHNT3qquVveWVlJRWHAye4v3lh4VQcemhzxt3lrKswB1TVlF91PPGCUeES/O0JJqnJZHV1P1s0SSRFk2SadHE7ctMS7JbaqZO4QyOQqB9wOdcb9tBqroZ1+JARJoC16lqsbfvAeABb9+zwFpgB5ApIgleq+OIcxrTmMXFCc1SEmmWknjY/3mOR1lFlbuVdpSQKTm4r5wde8tZt30vu/aVs6+87r/JRCAzNZHmTZJokZZ0cKqa5k2SyGpSEzYBn5sk0STJBgzEgnAGx0KgmzcKqhAYzqG+CQBEJBvYqarVwFjcCKuajvVMVd0hIj2BnsA7qqoiMgu4Hjey6hbg1TBegzGNRkpiPCmJ8bRslhLU98oqqthVWs7OfeXs2lfBjn0H2LWvnJ2lFe6n99q4s5TFm4rZta+cyqPcT0tKiKNFmguRmpZLTajUBM+h/S58bOblxidswaGqlSIyGpiBG447SVVXiMj9QL6qvgZcDPxBRBR3q+ou7+uJwBzvL5fduGG6Nf0aPwemisjvgMXAxHBdgzGxICUx3q3NknF8/Tiqyp4DlezcW87O0vLDwqX2583Fu9mx9wC7y44+WCA9OeGwYKnrttmh22lJNEtJtAc6fSaq0X/7Py8vT/Pz8/0uw5iYVVFVTXFpBbu822Q1LZyaV+DnXfvK2bGv/Ih1Y2rExwnNa/poAm6Ttajjc4um7mdqko1GOxEiskhV82pv97tz3BgTAxLj48hJTyYnPRlaHd93SssrD799VlrOzn0VB4Nll9fC+czrq9l1jBFpKYlxtExP4cozW3Nj3w50zDr6YmimfhYcxpiIVDMUOrf58R1fXa2U7K846u2z9UX7mDD3C/4xez0Du2VzU7+OXNa9pU3pfwIsOIwxUSEuTmju3aYip+5jtu0uY+qCTTy3YCN3/nsRrZulMLxPe4b37kDrjOAGFcQy6+MwxsScyqpq3l+9nWfmb2T22iLi44RBp7fkpn4duaBLtnW+e6yPwxhjPAnxcVxxRmuuOKM1G3bs49kFG3kxv4AZK7bRKSuNG/t24JvntXetF3MEa3EYYwxugbG3P93KM/M2sPDLXSQlxHHVWW24qV8Hzu3QPCYfbDxai8OCwxhjalm9dTfPzt/Iy58UsvdAJd1bpzOyX0eG9mpH0+TYuVFjwWHBYYwJ0r4Dlby2dDPPzNvAis27aZIUz7W92nFT3470aNvM7/LCzoLDgsMYc4JUlSWbipkyfyOvL93Mgcpqzu2Qyci+HbmqZ5uone7egsOCwxgTAsWl5UxbVMCz8zey/qt9ZKYlcv25uYzs15HO2dH1YKEFhwWHMSaEVJWPP9/BlPkbmbFiK5XVyoCu2Yzs24FBPVqRGAUPFlpwWHAYY8Jk++4ynl/oHizcXFJGy/Rkhvduz/A+HYJeBCySWHBYcBhjwqyqWpm1ejtT5m/gg7VFCHDZ6a0Y2bcDF3bLaXQPFtoDgMYYE2bxccKgHq0Y1KMVm3aW8uyCjbywcBMzV26jfYtUbuzTkRvycskKcmXISGMtDmOMCaMDlVXMWLGNZ+ZtYMEXO0mKj+PKs1ozsm9HeneK7AcL7VaVBYcxxmefbdvDlPkbeWlRAXsOVHJaq3RG9uvAkF7taJaS6Hd5R7DgsOAwxkSI0vJKXl+6mWfmbWR5YQlpSfFce05bRvbtyJntMvwu7yALDgsOY0wEWrqpmCnzN/Da0s2UVVRzdvtMburbgat7tvV95UJfgkNEBgN/wa05PkFV/1hrf0dgEm72/J24tcULvH1/Aq4C4oCZwI9UVUXkA6ANsN87zRWquv1YdVhwGGMiXUlpBS8vLuCZeRv4vGgfzVISuP689ozs14EuOU19qanBg0NE4oG1wOVAAbAQGKGqKwOOeRF4Q1Uni8ilwG2qerOInA88BFzoHToXGKuqH3jBMUZVjzsJLDiMMY2FqjJv/U6mzN/AjBVbqahS+p+SxU39OnJ5j1YkJTTcg4V+DMftA6xT1fVeAVOBa4GVAcf0AO7x3s8CXvHeK5ACJAECJALbwlirMcZEBBGhf5cs+nfJomjPAV7I38Sz8zdy17OfkN3UPVg4om8H2vn4YGE4o6sdsCngc4G3LdBSYJj3fiiQLiJZqvoxLki2eK8Zqroq4Hv/EpElIvK/cpSxbCIySkTyRSS/qKgoFNdjjDENKic9mbsu6crsn13CpFvzODs3gyc+WMfAB9/njqcWMmv1dqqqG76f2u8HAMcAj4vIrcBsoBCoEpGuwOlArnfcTBEZqKpzgJGqWigi6cBLwM3A07VPrKrjgfHgblWF/UqMMSZM4uOES7u34tLurSjYVcpzCzby/MJNvPfUdnKbpzKiTwduyGtPTnrDPFgYzhZHIdA+4HOut+0gVd2sqsNUtRdwr7etGNf6mKeqe1V1L/AW0N/bX+j93AM8i7slZowxMSG3eRo//Vp3PvrFZTx+Yy/aN0/joRlrOP+P7zH62U+Yt34H4R4tG84Wx0Kgm4h0xgXGcODGwANEJBvYqarVwFjcCCuAjcB3ReQPuD6Oi4A/i0gCkKmqX4lIInA18G4Yr8EYYyJSUkIcV/dsy9U927Ju+16mzN/AS4sKeGPZFrq2bMrIvh0Ydm4uGamhf7AwbC0OVa0ERgMzgFXAC6q6QkTuF5FrvMMuBtaIyFqgFfCAt30a8DmwHNcPslRVXweSgRkisgxYggukf4brGowxpjHo2rIp933jDOb/chB/ur4nTZLi+c3rK+n7+3dZXlAS8t9nDwAaY0wUWl5QwvTFhYz9evcTXhvEZsc1xpgYclZuBmflhmf6ksa/RJUxxpgGZcFhjDEmKBYcxhhjgmLBYYwxJigWHCa27d4MK6ZD4SKIgRGGxoSCjaoysWX3ZvhyLnw5x/3cuf7QvswOcMZQ92pzDkTwkp7G+MmCw0S3kkLY8OGRQZGSAR0vgN7fgfb9oGi1a3l8/AR8+Bdo3gl6DPFC5GwLEWMC2AOAJrrUFxSdBrhXqzMhro7V1Up3wur/uBBZ/wFoFTTvfKgl0vosCxETM2zpWAuO6HSyQXEspTth1euw8hVY/18XIi26wBleS6TVmRYiJqpZcFhwRIdwBsWx7NsBq193LZEvZoNWQ1bXQy2Rlj0sREzUseCw4Gic/AqKY9n3Fax6DVa84urSasg+9VCfSMvTLURMVLDgsOBoHEoKDx/1tOsLtz0lAzoOCAiKMxouKI5lb5EXItNdwGk1ZJ8W0BLp7neFxpwwCw4LjsjU2ILiWPZscyGy8lV3LSjknH6oTyTnNL8rNCYoFhwWHJEhmoLiWPZsdR3rK6bDho8Adf0gNS2R7G5+V2hMvSw4LDj8EStBcSy7txzqE9n4MaCuT6amTyS7q98VGlMnCw4LjoZRUgBffhjbQXEsuzfDSq9PZNM8t63VWYduZ2V18bc+YwJYcFhwhIcFxYkrKXT9ISumQ8ECt611TxciPYZYiBjfWXBYcISGBUV4lBQEhMhCt63N2a4V0mMItOjsb30mJvkSHCIyGPgLEA9MUNU/1trfEZgE5AA7gZtUtcDb9yfgKtwMvjOBH6mqish5wFNAKvBmzfZj1WHBcRKOGhSZtZ6jsKAImeKNh0KkcJHb1raX1ycyxM2jZUwDaPDgEJF4YC1wOVAALARGqOrKgGNeBN5Q1ckicilwm6reLCLnAw8BF3qHzgXGquoHIrIAuBuYjwuOv6rqW8eqxYIjCKU74bOZxxEUZ0Kczcofdrs2HAqRzZ+4bW3P9UZnDXEz+hoTJkcLjnDOjtsHWKeq670CpgLXAisDjukB3OO9nwW84r1XIAVIAgRIBLaJSBugmarO8875NDAEOGZwmONUVQETLnNPZ9cERZ9RFhR+at4RLrjbvXZ96UZmrZgOM//XvdrlHeoTyWzvd7UmRoQzONoBmwI+FwB9ax2zFBiGu501FEgXkSxV/VhEZgFbcMHxuKquEpE87zyB52xX1y8XkVHAKIAOHeyvsuOy7AUXGkPHw1nftKCINM07wYAfu9fOL9zkiyumwzu/cq/c3l6fyLWQket3tSaK+f0vwxjgIhFZDFwEFAJVItIVOB3IxQXDpSIyMJgTq+p4Vc1T1bycnJxQ1x19qqtg7iNu2vCeN1hoRLoWnWHAT+DO2fDDT+CyX0NlGcz4JTx6Bky8Aj7+mxu5ZUyIhbPFUQgEtp1zvW0HqepmXIsDEWkKXKeqxSLyXWCequ719r0F9Af+7Z3nqOc0J2jlK7BjHXzzKZugr7HJ6gID/8e9dnzuWiErXoEZY92rfb9DLZFmbfyu1kSBcP5ZuRDoJiKdRSQJGA68FniAiGSLSE0NY3EjrAA24loiCSKSiGuNrFLVLcBuEeknIgJ8G3g1jNcQG1RhziNuhtfTr/G7GnMysrrAhWPg+3NhdD5c8is4sAfe/jk8cjpMuhLm/8M9zW7MCQpbcKhqJTAamAGsAl5Q1RUicr+I1PzrdDGwRkTWAq2AB7zt04DPgeW4fpClqvq6t+8HwARgnXeMdYyfrLVvw7ZPYcA9NqQ2mmR3g4t+Cj/4CO5aCJf8EsqK4a2fuRB55Qewd7vfVZpGyB4AjHWqbiTVviJ3rzw+0e+KTLhtXw2L/+1aHompcNHPoe+d9t/eHOFow3GtBzTWrf/APWQ24Cf2D0esaNkdvvYA/GAetO8L79wLT14An7/vd2WmkbDgiHVzHob0NnDOSL8rMQ0tuyuMfBFGPA9V5fDvoTB1pHtexJhjsOCIZRvnuSfEz/8hJCT7XY3xgwicNti1Pi77tWt1PNEXZv0eykv9rs5EKAuOWDZ7HKRlwXm3+l2J8VtiihvOOzoful8N/30QnujjhvXGQD+oCY4FR6zavATWzYR+P4CkJn5XYyJFRju4fiLc+qab8fjFW+Dpa2D7Kr8rMxHEgiNWzXkYkjOgz3f9rsREok4XwKj/wtfHwZZlrvP8rZ/D/mK/KzMRwIIjFm1f7ZYy7TvK/VVpTF3iE9wfFncvhvNuccN3HzsXFk2G6mq/qzM+suCIRXMfgcQ06Pt9vysxjUFaC7j6Ubjzv252gdfvhgmXwqaFfldmfGLBEWt2rofl0yDvdmiS5Xc1pjFpczbc9hYMmwB7tsLEQTD9+7Bnm9+VmQZmwRFr5v4Z4hKg/2i/KzGNkQj0/CaMXugeGl3+Ijx2Hnz0GFSW+12daSAWHLGkpBCWPAu9brJZUs3JSU6HQf8Hd82Hjv3deiBPng/r3vO7MtMALDhiyUePgVbDBT/yuxITLbK6uKfPb3wBtAqeGQbP3egWmjJRy4IjVuwtgkVPwdnD3XKkxoTSqV/znj6/z81/9kRfeP939vR5lLLgiBXznnArxA34id+VmGiVkAwD74Ef5kOPa2D2Q/B4b/j0ZXv6PMocMzhEJF5E7hSR34rIBbX2/Sq8pZmQ2b8LFkyAM4a4NRqMCadmbeG6CXDb25DWHKbdBpO/AdtW+F2ZCZH6Whz/wK2+twP4q4g8ErBvWNiqMqE1fzyU74GBY/yuxMSSjv3d0+dXPeIWCvv7QHjzZ+4PGdOo1RccfVT1RlX9M9AXaCoiL4tIMmALUzcGB/bC/Cfh1Cuh9Zl+V2NiTVw89L7DLRKWdxss/KcbvrvoKaiu8rs6c4LqC46kmjeqWqmqo4AlwPtA03AWZkIkf5L7C+9Ca20YH6W1gKsedi2Q7NPg9R/BPy+FTQv8rsycgPqCI19EBgduUNX7gX8Bneo7uYgMFpE1IrJORH5Rx/6OIvKeiCwTkQ9EJNfbfomILAl4lYnIEG/fUyLyRcC+c473YmNOxX43BPeUiyH3iNUfjWl4bXrCbW/CdRPdeucTL4eX73RPoptGI2xrjotIPLAWuBwoABYCI1R1ZcAxLwJvqOpkEbkUuE1Vb651nhbAOiBXVUtF5CnvO9OOt5aYXXN8wT/hzTFwyxvQeaDf1RhzuAN73SzNHz8O8Ulw0c/c/GkJSfV/1zSIk1pz3AuBYPUB1qnqelUtB6YC19Y6pgfuthfArDr2A1wPvKWqNiA8GJXl8OFf3JrSnQb4XY0xR0puCoPuc89/dBoAM38NT/aHz971uzJTj3qDQ0TSgVdP4NztgE0Bnwu8bYGWcmh01lAgXURqz7w3HHiu1rYHvNtbj3od9XXVPUpE8kUkv6io6ATKb+SWPQ8lm+DCn7r5hYyJVFld4Mbn4cYX3fMeU66D50a4CTlNRKrvOY42wLvA+DD9/jHARSKyGDfstxA4ONTC+/1nATMCvjMW6A70BloAP6/rxKo6XlXzVDUvJycnTOVHqOoqmPuom8206yC/qzHm+Jx6hWt9DPoNfDEbnugH7/0Wyvf5XZmppb4Wxxzgj6r62gmcuxBoH/A519t2kKpuVtVhqtoLuNfbFrjE2A3AdFWtCPjOFnUO4Drp+5xAbdFtxXTY+blbQ9paG6YxSUiCAT92a5+fMQTmjPOePn/Jnj6PIPUFxy6OvL10vBYC3USks4gk4W45HRZAIpItIjU1jAUm1TrHCGrdpvJaIYiIAEOAT0+wvuhUXe06HLNPg+7f8LsaY05MszYwbDzcPgPSsmDa7fDU1bDV/u8eCeoLjouBK0XkrmBPrKqVwGjcbaZVwAuqukJE7heRawLOv0ZE1gKtgAdqvi8inXAtlv/WOvUUEVkOLAeygd8FW1tUW/sWbF/pWhtxNhWZaeQ69INRH8DVf3b/u/7HQPjPGCjd6XdlMa3e4bjeiKp/qOp3Gqak0IuZ4biq7qGq/Tth9CK3ZrQx0aJ0J8z6PeRPhJRMuOx/4dxb3NPpJixOeDiuqlY15tCIKZ+/D5s/gQt+bKFhok9aC7hqHNw5B1r2gDd+AuMvho3z/K4s5pzQvQwRiRORkaEuxpykOQ9Dels450a/KzEmfFqfCbe+AddPgtIdMOlr8PIo2L3F78piRn3DcZuJyFgReVxErhDnh8B63IgnEyk2fAQbPoQL7nbrIhgTzUTgzOvc2ucDx7iRhI/nwdw/Q+UBv6uLevW1OP4NnIbriP4O7unu64EhqlrXU97GL7PHQVq2u+drTKxIauL6Ou6aD50vhHfvg7/1h89m+l1ZVKsvOE5R1VtV9R+4obE9gK+p6pLwl2aOW+En8Pl70P8uSErzuxpjGl6LU2DEczDyJdcamXI9PPst2PG535VFpfqCI/DBuyqgQFXLwluSCdqchyElA3rbGAYT47oNgu9/DJf/Fr6cC3/rB+/+xk2oaEKmvuA4W0R2e689QM+a9yKyuyEKNPXYthJWvwF97oSUZn5XY4z/EpJcX98PF7l+kLmPuKfPl0+zp89D5JjBoarxqtrMe6WrakLAe/tXKhLMfQQSm0C/7/tdiTGRJb01DP073P4ONM2Bl+6A1++Gqkq/K2v07NHixmzH524On963uzHuxpgjdegL353lZlP45GmYOsJuXZ0kC47GbO6jEJcI/Uf7XYkxkS0uHi77NVz9KKx7FyZf7VYgNCfEgqOxKimApVPh3G+7Jrkxpn55t8PwZ2H7apgwCL5a53dFjZIFR2P14V8BhQt+5HclxjQup10Jt/7HrfMx8XLYtMDvihodC47GaO92+GQy9BwOme3rP94Yc7jc8+COdyA1EyZ/A1a94XdFjYoFR2P08eNQVQ4DfuJ3JcY0Xlld4I6Z0OoMeP4mWPBPvytqNCw4GpvSnbBwIpwxFLK7+l2NMY1bk2y45Q04dTC8OQZm3ucWQzPHZMHR2CwYD+V73dBCY8zJS0qDbz3jOs4//DNMH2UTJdbDFm1oTA7sgXlPwmlXuea1MSY04hPgqkcgoz289xvYs9WFSWqm35VFJGtxNCYLJ0JZMVxorQ1jQk4EBt4DQ8e7xaH+daUb9m6OENbgEJHBIrJGRNaJyC/q2N9RRN4TkWUi8oGI5HrbLxGRJQGvMhEZ4u3rLCLzvXM+LyJJ4byGiFGx33WKn3IJtDvP72qMiV5nfwtumgbFm2DC5bBthd8VRZywBYe3VvkTwJW46dhHiEiPWoeNA55W1Z7A/cAfAFR1lqqeo6rnAJcCpcA73nceBB5V1a7ALuCOcF1DRPnkadhXBBf+1O9KjIl+p1wMt78FKEwaDOv/63NBkSWcLY4+wDpVXa+q5cBUoPbiTz2A9733s+rYD27hqLdUtVREBBck07x9k4EhIa880lSWw4d/gQ79odMFfldjTGxofZYbrtusLTxzHSx70e+KIkY4g6MdsCngc4G3LdBSYJj3fiiQLiJZtY4ZDjznvc8CilW1ZnrLus4JgIiMEpF8EckvKio6wUuIEMumwu5Ct0SmMabhZLaH29+G9n3h5e+4pWltanbfO8fHABeJyGLgIqAQqKrZKSJtgLOAGcGeWFXHq2qequbl5OSEqt6GV1UJcx6BNudA18v8rsaY2JPaHG5+2a3t8e598OZPobqq/u9FsXAOxy0EAufDyPW2HaSqm/FaHCLSFLhOVYsDDrkBmK6qNSsR7gAyRSTBa3Uccc6os2I67PrCDQ0U8bsaY2JTQjIMm+BuW330GOzZAtdNgMRUvyvzRThbHAuBbt4oqCTcLafXAg8QkWwRqalhLDCp1jlGcOg2FaqquL6Q671NtwCvhqH2yFBd7ZaFzTndPbthjPFPXBxc8TsY/CCs/g9Mvgb27fC7Kl+ELTi8FsFo3G2mVcALqrpCRGLTrB4AAA+2SURBVO4XkWu8wy4G1ojIWqAV8EDN90WkE67FUns4w8+Be0RkHa7PY2K4rsF3a/4DRavcU+Jxft9VNMYA0O97cMNk2LLUza678wu/K2pwojHQ0ZOXl6f5+fl+lxEcVRh/MZSVwOh892SrMSZybJwHzw2HuAS48QVod67fFYWciCxS1bza2+3P2Ej1+XuwZYmbAddCw5jI06GfW888MRWeugrWvlP/d6KEBUekmj0OmrWDs0f4XYkx5mhyToU73oXsbq71sWiy3xU1CAuOSPTlh7DxY7e6X0JszKhiTKOV3sqtKNjlEnj9bpj1+6h/1sOCIxLNGQdNctx64saYyJecDiOmQq+b4L8Pwqt3QVVF/d9rpOzmeaQpXASfvw+DfhOzY8SNaZTiE+Gax93U7B/8wU3NfsNkFypRxlockWb2w5CSCb1jY+5GY6KKCFz8C7jmMVj/Afzr6y5AoowFRyTZtsI9u9H3e1H5V4oxMePcb8ONz8OOz93U7EVr/K4opCw4IsmchyGpKfS90+9KjDEnq9vlcNt/oLIMJl4BGz72u6KQseCIFDs+d/NS9b4D0lr4XY0xJhTa9oLvzHSDXZ6+FlZGxwxJFhyRYu4jEJ8E/Uf7XYkxJpSad4I73oG258ALt8C8J/2u6KRZcESC4o2wdKq7L9q0pd/VGGNCLa0FfPtV6H4VvP0LmHGvm8S0kbLgiAQf/hUQOP9uvysxxoRLYirc8DT0uRM+fhxeuh0qyvyu6oTYcxx+27PNrSd+9nC32pgxJnrFxcOVD0JGLsz8X9i7HYZPcYtFNSLW4vDbx49BdYWbzNAYE/1E4IK74bqJULAQJg2G4k31fy+CWHD4qXQnLJzklqTM6uJ3NcaYhnTW9XDTy7B7C0wYBFuW+V3RcbPg8NP8v0PFPhhwj9+VGGP80Hkg3P62u4X1r6+76YYaAQsOv5TtdsHR/Wpo1cPvaowxfmnVA77zLjTvCFO+6UZYRjgLDr8snOBW9xv4P35XYozxW7O2cNub0PECmH6nW48ngqdmt+DwQ3kpfPwEdLksKpebNMacgJQMGDkNzroB3v8t/OceqKr0u6o6hTU4RGSwiKwRkXUi8os69ncUkfdEZJmIfCAiuQH7OojIOyKySkRWikgnb/tTIvKFiCzxXueE8xrC4pPJUPoVXDjG70qMMZEkIQmGjXejLPMnwfM3Qfk+v6s6QtiCQ0TigSeAK4EewAgRqX0zfxzwtKr2BO4H/hCw72ngIVU9HegDbA/Y91NVPcd7LQnXNYRF5QH3wF/HC6Dj+X5XY4yJNCIw6P/g6+Pgsxkw+Ruwt8jvqg4TzhZHH2Cdqq5X1XJgKnBtrWN6ADXDCGbV7PcCJkFVZwKo6l5VLQ1jrQ1n6XOwZ7P1bRhjjq3Pd+Fbz8C2lTDxcjcRaoQIZ3C0AwKfainwtgVaCgzz3g8F0kUkCzgVKBaRl0VksYg85LVgajzg3d56VESS6/rlIjJKRPJFJL+oKELSuqoS5j7qZszscqnf1RhjIl33q+CW1+HAbhceBfl+VwT43zk+BrhIRBYDFwGFQBVuKpSB3v7ewCnArd53xgLdve0tgJ/XdWJVHa+qeaqal5OTE85rOH6fvgS7voQLf+qao8YYU5/2veGOmZDcDJ66Gla/6XdFYQ2OQiBw8qVcb9tBqrpZVYepai/gXm9bMa51ssS7zVUJvAKc6+3fos4B4F+4W2KRr7raTZ3esgeceqXf1RhjGpOsLi48Wp4Oz4+EhRN9LSecwbEQ6CYinUUkCRgOvBZ4gIhki0hNDWOBSQHfzRSRmqbCpcBK7zttvJ8CDAE+DeM1hM7qN6BotevbiPO7oWeMaXSa5sCtb0C3K9xQ3Xd/49uzHmH7F8xrKYwGZgCrgBdUdYWI3C8i13iHXQysEZG1QCvgAe+7VbjbVO+JyHJAgH9635nibVsOZAO/C9c1hIwqzH4IWnSBM4b6XY0xprFKagLfmgLn3eruYEz/HlSWN3gZYZ1WXVXfBN6ste3XAe+nAdOO8t2ZQM86tje+XuV178LWZXDN425OGmOMOVHxCXD1n93U7O//DvZuhRv+DSnNGqwEu2cSbjWtjWa50PNbfldjjIkGIm6QzZAn4cu58K8rYffmBvv1Fhzh9uVc2DQfBvzYPRVqjDGhcs6NMPJF2LUBJlwO21c1yK+14Ai3OeOgSUvodZPflRhjolGXS90EidWVMOlr7o/VMLPgCKeCfFj/AZw/2q03bIwx4dCmJ3xnJjRtDf8e6p4ZCyMLjnCaPc6tJZx3u9+VGGOiXWYHuGMGtMuDabfDR4+FbbiuBUe4bP0U1r4Ffb8Pyel+V2OMiQWpzeHm6dBjCLzzK3h7LFRXhfzXWHCEy5yHISkd+o7yuxJjTCxJTIHr/wX9R8OC8bA59BOIh/U5jpj11WewYrobSZXa3O9qjDGxJi4OvvYAnDMyLEtTW4sjHOY+Cgkp0O8uvysxxsSyMIQGWHCE3q4NsOx5OO8WN7eMMcZEGQuOUPvwL4DA+Xf7XYkxxoSFBUco7dkKi59xT3Nm1F6zyhhjooMFRyh99BhUV7hOcWOMiVIWHKGybwfkT4KzvgktTvG7GmOMCRsLjlCZ/yRUlMKAe/yuxBhjwsqCIxTKSmD+eDj9G9Cyu9/VGGNMWFlwhMKCf8KBErcsrDHGRDkLjpNVvg/m/Q26Xg5te/ldjTHGhF1Yg0NEBovIGhFZJyK/qGN/RxF5T0SWicgHIpIbsK+DiLwjIqtEZKWIdPK2dxaR+d45nxcRf1dHWjQZSnfAhWN8LcMYYxpK2IJDROKBJ4ArgR7ACBGp/fz7OOBpVe0J3A/8IWDf08BDqno60AfY7m1/EHhUVbsCu4A7wnUN9ao8AB/9FToOgA79fCvDGGMaUjhbHH2Adaq6XlXLganAtbWO6QG8772fVbPfC5gEVZ0JoKp7VbVURAS4FJjmfWcyMCSM13BsS6bAni3W2jDGxJRwBkc7YFPA5wJvW6ClwDDv/VAgXUSygFOBYhF5WUQWi8hDXgsmCyhW1cpjnBMAERklIvkikl9UVBSiSwpQVeEmM2x3HpxycejPb4wxEcrvzvExwEUishi4CCgEqnDTvQ/09vcGTgFuDebEqjpeVfNUNS8nJwyTDS6fBsUbYeAYEAn9+Y0xJkKFMzgKgfYBn3O9bQep6mZVHaaqvYB7vW3FuJbEEu82VyXwCnAusAPIFJGEo52zQVRXw9xHoNWZcOrgBv/1xhjjp3AGx0KgmzcKKgkYDrwWeICIZItITQ1jgUkB380UkZqmwqXASlVVXF/I9d72W4BXw3gNdVv1Gny1Fgbe4xZMMcaYGBK2f/W8lsJoYAawCnhBVVeIyP0ico132MXAGhFZC7QCHvC+W4W7TfWeiCwHBPin952fA/eIyDpcn8fEcF1DnVRh9jjI6urW9TXGmBgT1qVjVfVN4M1a234d8H4ah0ZI1f7uTKBnHdvX40Zs+eOzd2Dbcrj2bxAX71sZxhjjF7vPEgxVmP0QZHSAnjf4XY0xxvjCgiMYX8yGgoVwwd0Qn+h3NcYY4wsLjmDMGQdNW0Gvm/2uxBhjfGPBcbw2LXQtjvN/CIkpfldjjDG+seA4XnPGQWpzOO82vysxxhhfWXAcjy3LYO3b0O8uSG7qdzXGGOMrC47jMedhSG4Gfb7rdyXGGOM7C476FK2Fla9C7+9Aaqbf1RhjjO8sOOoz9xFISIH+d/ldiTHGRAQLjmPZ9SUsewHyboMm2X5XY4wxEcGC41g+/IubVuT8H/pdiTHGRAwLjmPJ7OhuUTVr63clxhgTMcI6yWGjN+DHfldgjDERx1ocxhhjgmLBYYwxJigWHMYYY4JiwWGMMSYoFhzGGGOCYsFhjDEmKBYcxhhjgmLBYYwxJiiiqn7XEHYiUgRsOMGvZwNfhbCcxsCuOTbYNUe/k73ejqqaU3tjTATHyRCRfFXN87uOhmTXHBvsmqNfuK7XblUZY4wJigWHMcaYoFhw1G+83wX4wK45Ntg1R7+wXK/1cRhjjAmKtTiMMcYExYLDGGNMUCw4joOI/FZElonIEhF5R0SifklAEXlIRFZ71z1dRDL9rincROSbIrJCRKpFJGqHbIrIYBFZIyLrROQXftcTbiIySUS2i8inftfSUESkvYjMEpGV3v+mfxTK81twHJ+HVLWnqp4DvAH82u+CGsBM4ExV7QmsBcb6XE9D+BQYBsz2u5BwEZF44AngSqAHMEJEevhbVdg9BQz2u4gGVgn8j6r2APoBd4Xyv7MFx3FQ1d0BH5sAUT+iQFXfUdVK7+M8INfPehqCqq5S1TV+1xFmfYB1qrpeVcuBqcC1PtcUVqo6G9jpdx0NSVW3qOon3vs9wCqgXajOb2uOHycReQD4NlACXOJzOQ3tduB5v4swIdEO2BTwuQDo61MtpgGISCegFzA/VOe04PCIyLtA6zp23auqr6rqvcC9IjIWGA3c16AFhkF91+wdcy+u2TulIWsLl+O5ZmOihYg0BV4CflzrzslJseDwqOqg4zx0CvAmURAc9V2ziNwKXA1cplHywE8Q/52jVSHQPuBzrrfNRBkRScSFxhRVfTmU57Y+juMgIt0CPl4LrParloYiIoOBnwHXqGqp3/WYkFkIdBORziKSBAwHXvO5JhNiIiLARGCVqj4S8vNHyR+SYSUiLwGnAdW46dm/p6pR/VeaiKwDkoEd3qZ5qvo9H0sKOxEZCjwG5ADFwBJV/Zq/VYWeiHwd+DMQD0xS1Qd8LimsROQ54GLcFOPbgPtUdaKvRYWZiAwA5gDLcf9uAfxSVd8MyfktOIwxxgTDblUZY4wJigWHMcaYoFhwGGOMCYoFhzHGmKBYcBhjjAmKBYcxYSYiX4pI9skeY0yksOAwxhgTFAsOY0JIRF4RkUXeGgijau3r5K1xMkVEVonINBFJCzjkhyLyiYgsF5Hu3nf6iMjHIrJYRD4SkdMa9IKMqYMFhzGhdbuqngfkAXeLSFat/acBf1PV04HdwA8C9n2lqucCTwJjvG2rgYGq2gu3Dszvw1q9McfBgsOY0LpbRJbi1jBpD3SrtX+Tqn7ovX8GGBCwr2YiukVAJ+99BvCit3rdo8AZ4SjamGBYcBgTIiJyMTAI6K+qZwOLgZRah9We4yfw8wHvZxWHZq7+LTBLVc8EvlHH+YxpcBYcxoROBrBLVUu9Pop+dRzTQUT6e+9vBOYexzlrJtS8NSRVGnOSLDiMCZ23gQQRWQX8EXe7qrY1uPWfVwHNcf0Zx/In4A8ishhbP8dECJsd15gG4i3h+YZ328mYRstaHMYYY4JiLQ5jjDFBsRaHMcaYoFhwGGOMCYoFhzHGmKBYcBhjjAmKBYcxxpig/D8KCYvOWD0v8wAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GkMaLvChUiV0",
        "outputId": "4768bb5d-713d-4aaa-be53-59f7b66a158b"
      },
      "source": [
        "ridge = Ridge(alpha=0.1)\n",
        "ridge.fit(train_scaled, train_target)\n",
        "\n",
        "print(ridge.score(train_scaled, train_target))\n",
        "print(ridge.score(test_scaled, test_target))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9903815817570366\n",
            "0.9827976465386927\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pWyXspRIT9NT"
      },
      "source": [
        "### 라쏘 회귀\n",
        "과대적합 감소"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YOFWtLa2Ujsl",
        "outputId": "48bfb0d9-0546-4e95-f107-1b20de29289d"
      },
      "source": [
        "from sklearn.linear_model import Lasso\n",
        "\n",
        "lasso = Lasso()\n",
        "lasso.fit(train_scaled, train_target)\n",
        "print(lasso.score(train_scaled, train_target))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9897898972080961\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lwyTARNOUk7q",
        "outputId": "6e615b31-b3db-4f8b-a45a-52e005ab1bf0"
      },
      "source": [
        "print(lasso.score(test_scaled, test_target))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9800593698421883\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pkhj41XoUn-Z",
        "outputId": "2f8e5e04-1a2e-44d4-8c58-00d816c1ce97"
      },
      "source": [
        "train_score = []\n",
        "test_score = []\n",
        "\n",
        "alpha_list = [0.001, 0.01, 0.1, 1, 10, 100]\n",
        "for alpha in alpha_list:\n",
        "    # 라쏘 모델을 만듭니다\n",
        "    lasso = Lasso(alpha=alpha, max_iter=10000)\n",
        "    # 라쏘 모델을 훈련합니다\n",
        "    lasso.fit(train_scaled, train_target)\n",
        "    # 훈련 점수와 테스트 점수를 저장합니다\n",
        "    train_score.append(lasso.score(train_scaled, train_target))\n",
        "    test_score.append(lasso.score(test_scaled, test_target))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_coordinate_descent.py:476: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations. Duality gap: 18778.697957792876, tolerance: 518.2793833333334\n",
            "  positive)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_coordinate_descent.py:476: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations. Duality gap: 12972.821345404844, tolerance: 518.2793833333334\n",
            "  positive)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "2OqIUdtIUpoF",
        "outputId": "ca20b19c-754c-448a-bd28-8dc4cf67268b"
      },
      "source": [
        "plt.plot(np.log10(alpha_list), train_score)\n",
        "plt.plot(np.log10(alpha_list), test_score)\n",
        "plt.xlabel('alpha')\n",
        "plt.ylabel('R^2')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-UedFO6JUrGC",
        "outputId": "21b1f9e8-4bef-474f-9ba1-254e7c7efb8a"
      },
      "source": [
        "lasso = Lasso(alpha=10)\n",
        "lasso.fit(train_scaled, train_target)\n",
        "\n",
        "print(lasso.score(train_scaled, train_target))\n",
        "print(lasso.score(test_scaled, test_target))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "0.9888067471131867\n",
            "0.9824470598706695\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "O_1HBKhpUsOL",
        "outputId": "83ff4390-af8c-414b-f925-d14c0bc0135b"
      },
      "source": [
        "print(np.sum(lasso.coef_ == 0))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "40\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}