ch02/ch02.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "JmzQnrj_qAK2"
   },
   "source": [
    "# 머신 러닝 교과서 3판"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "OrbmAIqKqAK6"
   },
   "source": [
    "# 2장 - 분류를 위한 머신러닝 알고리즘 훈련"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "6mi4HcBkqAK7"
   },
   "source": [
    "**아래 링크를 통해 이 노트북을 주피터 노트북 뷰어(nbviewer.jupyter.org)로 보거나 구글 코랩(colab.research.google.com)에서 실행할 수 있습니다.**\n",
    "\n",
    "<table class=\"tfo-notebook-buttons\" align=\"left\">\n",
    "  <td>\n",
    "    <a target=\"_blank\" href=\"https://nbviewer.jupyter.org/github/rickiepark/python-machine-learning-book-3rd-edition/blob/master/ch02/ch02.ipynb\"><img src=\"https://jupyter.org/assets/main-logo.svg\" width=\"28\" />주피터 노트북 뷰어로 보기</a>\n",
    "  </td>\n",
    "  <td>\n",
    "    <a target=\"_blank\" href=\"https://colab.research.google.com/github/rickiepark/python-machine-learning-book-3rd-edition/blob/master/ch02/ch02.ipynb\"><img src=\"https://www.tensorflow.org/images/colab_logo_32px.png\" />구글 코랩(Colab)에서 실행하기</a>\n",
    "  </td>\n",
    "</table>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "H6ohB3qhqAK7"
   },
   "source": [
    "### 목차\n",
    "\n",
    "- 인공 뉴런: 초기 머신 러닝의 간단한 역사\n",
    "    - 인공 뉴런의 수학적 정의\n",
    "    - 퍼셉트론 학습 규칙\n",
    "- 파이썬으로 퍼셉트론 학습 알고리즘 구현\n",
    "    - 객체 지향 퍼셉트론 API\n",
    "    - 붓꽃 데이터셋에서 퍼셉트론 훈련\n",
    "- 적응형 선형 뉴런과 학습의 수렴\n",
    "    - 경사 하강법으로 비용 함수 최소화\n",
    "    - 파이썬으로 아달린 구현\n",
    "    - 특성 스케일을 조정하여 경사 하강법 결과 향상\n",
    "    - 대규모 머신 러닝과 확률적 경사 하강법\n",
    "- 요약"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hjRNovxBqAK8"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:52.878955Z",
     "iopub.status.busy": "2020-10-13T15:10:52.877244Z",
     "iopub.status.idle": "2020-10-13T15:10:52.880800Z",
     "shell.execute_reply": "2020-10-13T15:10:52.881676Z"
    },
    "id": "H7hW_HzPqAK8"
   },
   "outputs": [],
   "source": [
    "from IPython.display import Image"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Fr9HSqsoqAK9"
   },
   "source": [
    "# 인공 뉴런: 초기 머신 러닝의 간단한 역사"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 235
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:52.887495Z",
     "iopub.status.busy": "2020-10-13T15:10:52.886536Z",
     "iopub.status.idle": "2020-10-13T15:10:53.030753Z",
     "shell.execute_reply": "2020-10-13T15:10:53.031953Z"
    },
    "id": "-q5FhqX_qAK9",
    "outputId": "14b3199e-c905-4dde-e34d-0e1ed65bcbe0"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<img src=\"https://git.io/JtIbf\" width=\"500\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 2,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(url='https://git.io/JtIbf', width=500) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "qcS1fpThqAK-"
   },
   "source": [
    "## 인공 뉴런의 수학적 정의"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 301
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:53.039926Z",
     "iopub.status.busy": "2020-10-13T15:10:53.038740Z",
     "iopub.status.idle": "2020-10-13T15:10:53.078763Z",
     "shell.execute_reply": "2020-10-13T15:10:53.079671Z"
    },
    "id": "m2egC3TBqAK-",
    "outputId": "2c686553-4845-4a9e-f165-08de430b7cd5"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<img src=\"https://git.io/JtIbL\" width=\"500\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 3,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(url='https://git.io/JtIbL', width=500) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "q4CbGXjyqAK-"
   },
   "source": [
    "## 퍼셉트론 학습 규칙"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "y1smPkjyqAK_"
   },
   "source": [
    "$w_j := w_j + \\Delta w_j$\n",
    "\n",
    "$\\Delta w_j = \\eta(y^{(i)}-\\hat{y}^{(i)})x_j^{(i)}$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "XApIteAcqAK_"
   },
   "source": [
    "$y^{(i)}=-1, \\quad \\hat{y}^{(i)}=-1, \\qquad \\Delta w_j=\\eta(-1-(-1))x_j^{(i)}=0$\n",
    "\n",
    "$y^{(i)}=1, \\quad \\hat{y}^{(i)}=1, \\qquad \\Delta w_j=\\eta(1-1))x_j^{(i)}=0$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "u55ceXR9qAK_"
   },
   "source": [
    "$y^{(i)}=1, \\quad \\hat{y}^{(i)}=-1, \\qquad \\Delta w_j=\\eta(1-(-1))x_j^{(i)}=\\eta(2)x_j^{(i)}$\n",
    "\n",
    "$y^{(i)}=-1, \\quad \\hat{y}^{(i)}=1, \\qquad \\Delta w_j=\\eta(-1-1))x_j^{(i)}=\\eta(-2)x_j^{(i)}$"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "j6jDxcEoqAK_"
   },
   "source": [
    "$\\hat{y}^{(i)}=-1, \\quad y^{(i)}=+1, \\quad \\eta=1$\n",
    "\n",
    "$\\Delta w_j=(1-(-1))0.5=(2)0.5=1$\n",
    "\n",
    "$\\Delta w_j=(1-(-1))2=(2)2=4$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 191
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:53.088620Z",
     "iopub.status.busy": "2020-10-13T15:10:53.087535Z",
     "iopub.status.idle": "2020-10-13T15:10:53.126194Z",
     "shell.execute_reply": "2020-10-13T15:10:53.127170Z"
    },
    "id": "QeJHXlU4qAK_",
    "outputId": "c7626065-f093-4a77-b1be-a3e4397701d8"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<img src=\"https://git.io/JtIbq\" width=\"600\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 4,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(url='https://git.io/JtIbq', width=600) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 259
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:53.134460Z",
     "iopub.status.busy": "2020-10-13T15:10:53.133301Z",
     "iopub.status.idle": "2020-10-13T15:10:53.150276Z",
     "shell.execute_reply": "2020-10-13T15:10:53.151377Z"
    },
    "id": "kRRcFM_BqALA",
    "outputId": "4db20a88-ff03-4c84-db79-93899a247d38"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<img src=\"https://git.io/JtIbO\" width=\"600\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 5,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(url='https://git.io/JtIbO', width=600) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "7g3fMsmaqALA"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "E813AMpDqALA"
   },
   "source": [
    "# 파이썬으로 퍼셉트론 학습 알고리즘 구현"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "X53-_PH3qALA"
   },
   "source": [
    "## 객체 지향 퍼셉트론 API"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:53.176262Z",
     "iopub.status.busy": "2020-10-13T15:10:53.174446Z",
     "iopub.status.idle": "2020-10-13T15:10:53.178132Z",
     "shell.execute_reply": "2020-10-13T15:10:53.179538Z"
    },
    "id": "UYXKL2ixqALA"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "\n",
    "\n",
    "class Perceptron(object):\n",
    "    \"\"\"퍼셉트론 분류기\n",
    "\n",
    "    매개변수\n",
    "    ------------\n",
    "    eta : float\n",
    "      학습률 (0.0과 1.0 사이)\n",
    "    n_iter : int\n",
    "      훈련 데이터셋 반복 횟수\n",
    "    random_state : int\n",
    "      가중치 무작위 초기화를 위한 난수 생성기 시드\n",
    "\n",
    "    속성\n",
    "    -----------\n",
    "    w_ : 1d-array\n",
    "      학습된 가중치\n",
    "    errors_ : list\n",
    "      에포크마다 누적된 분류 오류\n",
    "\n",
    "    \"\"\"\n",
    "    def __init__(self, eta=0.01, n_iter=50, random_state=1):\n",
    "        self.eta = eta\n",
    "        self.n_iter = n_iter\n",
    "        self.random_state = random_state\n",
    "\n",
    "    def fit(self, X, y):\n",
    "        \"\"\"훈련 데이터 학습\n",
    "\n",
    "        매개변수\n",
    "        ----------\n",
    "        X : {array-like}, shape = [n_samples, n_features]\n",
    "          n_samples개의 샘플과 n_features개의 특성으로 이루어진 훈련 데이터\n",
    "        y : array-like, shape = [n_samples]\n",
    "          타깃값\n",
    "\n",
    "        반환값\n",
    "        -------\n",
    "        self : object\n",
    "\n",
    "        \"\"\"\n",
    "        rgen = np.random.RandomState(self.random_state)\n",
    "        self.w_ = rgen.normal(loc=0.0, scale=0.01, size=1 + X.shape[1])\n",
    "        self.errors_ = []\n",
    "\n",
    "        for _ in range(self.n_iter):\n",
    "            errors = 0\n",
    "            for xi, target in zip(X, y):\n",
    "                update = self.eta * (target - self.predict(xi))\n",
    "                self.w_[1:] += update * xi\n",
    "                self.w_[0] += update\n",
    "                errors += int(update != 0.0)\n",
    "            self.errors_.append(errors)\n",
    "        return self\n",
    "\n",
    "    def net_input(self, X):\n",
    "        \"\"\"입력 계산\"\"\"\n",
    "        return np.dot(X, self.w_[1:]) + self.w_[0]\n",
    "\n",
    "    def predict(self, X):\n",
    "        \"\"\"단위 계단 함수를 사용하여 클래스 레이블을 반환합니다\"\"\"\n",
    "        return np.where(self.net_input(X) >= 0.0, 1, -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:53.188993Z",
     "iopub.status.busy": "2020-10-13T15:10:53.187948Z",
     "iopub.status.idle": "2020-10-13T15:10:53.192416Z",
     "shell.execute_reply": "2020-10-13T15:10:53.193272Z"
    },
    "id": "Wi9_qzykqALB",
    "outputId": "84709a58-7368-4cb0-d039-c577614e4a42"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.0"
      ]
     },
     "execution_count": 7,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "v1 = np.array([1, 2, 3])\n",
    "v2 = 0.5 * v1\n",
    "np.arccos(v1.dot(v2) / (np.linalg.norm(v1) * np.linalg.norm(v2)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "9QyeWutEqALB"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "TXzkl6OdqALB"
   },
   "source": [
    "## 붓꽃 데이터셋에서 퍼셉트론 훈련"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "mD2Tj5BHqALB"
   },
   "source": [
    "..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "raRwqKBRqALC"
   },
   "source": [
    "### 붓꽃 데이터셋 읽기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 215
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:53.202418Z",
     "iopub.status.busy": "2020-10-13T15:10:53.201325Z",
     "iopub.status.idle": "2020-10-13T15:10:54.802200Z",
     "shell.execute_reply": "2020-10-13T15:10:54.803629Z"
    },
    "id": "7fZajroDqALC",
    "outputId": "4fd99a86-425b-4e39-f844-860ed73944a2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "URL: https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>145</th>\n",
       "      <td>6.7</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.2</td>\n",
       "      <td>2.3</td>\n",
       "      <td>Iris-virginica</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146</th>\n",
       "      <td>6.3</td>\n",
       "      <td>2.5</td>\n",
       "      <td>5.0</td>\n",
       "      <td>1.9</td>\n",
       "      <td>Iris-virginica</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>147</th>\n",
       "      <td>6.5</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.2</td>\n",
       "      <td>2.0</td>\n",
       "      <td>Iris-virginica</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>148</th>\n",
       "      <td>6.2</td>\n",
       "      <td>3.4</td>\n",
       "      <td>5.4</td>\n",
       "      <td>2.3</td>\n",
       "      <td>Iris-virginica</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149</th>\n",
       "      <td>5.9</td>\n",
       "      <td>3.0</td>\n",
       "      <td>5.1</td>\n",
       "      <td>1.8</td>\n",
       "      <td>Iris-virginica</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       0    1    2    3               4\n",
       "145  6.7  3.0  5.2  2.3  Iris-virginica\n",
       "146  6.3  2.5  5.0  1.9  Iris-virginica\n",
       "147  6.5  3.0  5.2  2.0  Iris-virginica\n",
       "148  6.2  3.4  5.4  2.3  Iris-virginica\n",
       "149  5.9  3.0  5.1  1.8  Iris-virginica"
      ]
     },
     "execution_count": 8,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "\n",
    "\n",
    "s = 'https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data'\n",
    "print('URL:', s)\n",
    "\n",
    "df = pd.read_csv(s,\n",
    "                 header=None,\n",
    "                 encoding='utf-8')\n",
    "\n",
    "df.tail()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "YFUKi7gUqALC"
   },
   "source": [
    "<hr>\n",
    "\n",
    "### 노트:\n",
    "\n",
    "이 깃허브에는 붓꽃 데이터셋이 포함되어 있습니다(이 책에서 사용하는 다른 데이터셋도 모두 포함되어 있습니다). 인터넷에 연결되어 있지 않거나 이따금 UCI 서버(https://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data)에 접근할 수 없을 때 사용할 수 있습니다. 이런 경우 로컬 디렉터리에서 붓꽃 데이터를 로드하려면 다음 코드를\n",
    "```python\n",
    "    df = pd.read_csv('https://archive.ics.uci.edu/ml/'\n",
    "        'machine-learning-databases/iris/iris.data', header=None)\n",
    "```\n",
    "다음과 같이 바꿉니다.\n",
    "```python \n",
    "    df = pd.read_csv('your/local/path/to/iris.data', header=None)\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "v_YynBiqqALC"
   },
   "source": [
    "그다음 다음 코드를 실행합니다.\n",
    "```python\n",
    "    df = pd.read_csv('iris.data', header=None, encoding='utf-8')\n",
    "    df.tail()\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "EI4UZuxgqALC"
   },
   "source": [
    "<hr>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "-ZP9wCdhqALD"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Lc6Bt3ogqALD"
   },
   "source": [
    "### 붓꽃 데이터 그래프 그리기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:54.953484Z",
     "iopub.status.busy": "2020-10-13T15:10:54.952117Z",
     "iopub.status.idle": "2020-10-13T15:10:55.624499Z",
     "shell.execute_reply": "2020-10-13T15:10:55.623986Z"
    },
    "id": "XcshxMesqALD",
    "outputId": "6f8d8f17-cc6b-4888-cbd3-253ab77b876e"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "\n",
    "# setosa와 versicolor를 선택합니다\n",
    "y = df.iloc[0:100, 4].values\n",
    "y = np.where(y == 'Iris-setosa', -1, 1)\n",
    "\n",
    "# 꽃받침 길이와 꽃잎 길이를 추출합니다\n",
    "X = df.iloc[0:100, [0, 2]].values\n",
    "\n",
    "# 산점도를 그립니다\n",
    "plt.scatter(X[:50, 0], X[:50, 1],\n",
    "            color='red', marker='o', label='setosa')\n",
    "plt.scatter(X[50:100, 0], X[50:100, 1],\n",
    "            color='blue', marker='x', label='versicolor')\n",
    "\n",
    "plt.xlabel('sepal length [cm]')\n",
    "plt.ylabel('petal length [cm]')\n",
    "plt.legend(loc='upper left')\n",
    "\n",
    "# plt.savefig('images/02_06.png', dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "q_dXznKqqALD"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "i1Iv80F_qALD"
   },
   "source": [
    "### 퍼셉트론 모델 훈련하기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:55.689301Z",
     "iopub.status.busy": "2020-10-13T15:10:55.636285Z",
     "iopub.status.idle": "2020-10-13T15:10:55.996833Z",
     "shell.execute_reply": "2020-10-13T15:10:55.995450Z"
    },
    "id": "i7rZoVY9qALE",
    "outputId": "0a368139-14e7-4260-9337-4f643406d0f4"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ppn = Perceptron(eta=0.1, n_iter=10)\n",
    "\n",
    "ppn.fit(X, y)\n",
    "\n",
    "plt.plot(range(1, len(ppn.errors_) + 1), ppn.errors_, marker='o')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Number of updates')\n",
    "\n",
    "# plt.savefig('images/02_07.png', dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "0j7_1-RTqALE"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "uPBaThajqALE"
   },
   "source": [
    "### 결정 경계 그래프 함수"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:56.017335Z",
     "iopub.status.busy": "2020-10-13T15:10:56.015706Z",
     "iopub.status.idle": "2020-10-13T15:10:56.020305Z",
     "shell.execute_reply": "2020-10-13T15:10:56.018987Z"
    },
    "id": "RKhIj0ynqALE"
   },
   "outputs": [],
   "source": [
    "from matplotlib.colors import ListedColormap\n",
    "\n",
    "\n",
    "def plot_decision_regions(X, y, classifier, resolution=0.02):\n",
    "\n",
    "    # 마커와 컬러맵을 설정합니다\n",
    "    markers = ('s', 'x', 'o', '^', 'v')\n",
    "    colors = ('red', 'blue', 'lightgreen', 'gray', 'cyan')\n",
    "    cmap = ListedColormap(colors[:len(np.unique(y))])\n",
    "\n",
    "    # 결정 경계를 그립니다\n",
    "    x1_min, x1_max = X[:, 0].min() - 1, X[:, 0].max() + 1\n",
    "    x2_min, x2_max = X[:, 1].min() - 1, X[:, 1].max() + 1\n",
    "    xx1, xx2 = np.meshgrid(np.arange(x1_min, x1_max, resolution),\n",
    "                           np.arange(x2_min, x2_max, resolution))\n",
    "    Z = classifier.predict(np.array([xx1.ravel(), xx2.ravel()]).T)\n",
    "    Z = Z.reshape(xx1.shape)\n",
    "    plt.contourf(xx1, xx2, Z, alpha=0.3, cmap=cmap)\n",
    "    plt.xlim(xx1.min(), xx1.max())\n",
    "    plt.ylim(xx2.min(), xx2.max())\n",
    "\n",
    "    # 샘플의 산점도를 그립니다\n",
    "    for idx, cl in enumerate(np.unique(y)):\n",
    "        plt.scatter(x=X[y == cl, 0], \n",
    "                    y=X[y == cl, 1],\n",
    "                    alpha=0.8, \n",
    "                    c=colors[idx],\n",
    "                    marker=markers[idx], \n",
    "                    label=cl, \n",
    "                    edgecolor='black')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 280
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:56.027421Z",
     "iopub.status.busy": "2020-10-13T15:10:56.026099Z",
     "iopub.status.idle": "2020-10-13T15:10:56.456714Z",
     "shell.execute_reply": "2020-10-13T15:10:56.456194Z"
    },
    "id": "ux8HGXQSqALE",
    "outputId": "7b6bb31b-07a0-4cb9-c681-055e601bbf6f"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_decision_regions(X, y, classifier=ppn)\n",
    "plt.xlabel('sepal length [cm]')\n",
    "plt.ylabel('petal length [cm]')\n",
    "plt.legend(loc='upper left')\n",
    "\n",
    "# plt.savefig('images/02_08.png', dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "WEdYJwaXqALF"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "VwzhuiRWqALF"
   },
   "source": [
    "# 적응형 선형 뉴런과 학습의 수렴"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "jtCbwYz7qALF"
   },
   "source": [
    "..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "GhERmxUDqALF"
   },
   "source": [
    "## 경사 하강법으로 비용 함수 최소화"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 451
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:56.464540Z",
     "iopub.status.busy": "2020-10-13T15:10:56.463527Z",
     "iopub.status.idle": "2020-10-13T15:10:56.520172Z",
     "shell.execute_reply": "2020-10-13T15:10:56.520593Z"
    },
    "id": "H3SPVRtIqALF",
    "outputId": "dd114d1e-2612-41a5-8b28-57c810c039c7"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<img src=\"https://git.io/JtIbn\" width=\"600\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 13,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(url='https://git.io/JtIbn', width=600) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 275
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:56.527815Z",
     "iopub.status.busy": "2020-10-13T15:10:56.526455Z",
     "iopub.status.idle": "2020-10-13T15:10:56.560118Z",
     "shell.execute_reply": "2020-10-13T15:10:56.560903Z"
    },
    "id": "tJS1Uk5ZqALF",
    "outputId": "9c1aaa61-5ecb-4384-c528-d84c51737159"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<img src=\"https://git.io/JtIbc\" width=\"500\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 14,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(url='https://git.io/JtIbc', width=500)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hmv7txAcqALG"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "tJ0OEDXpqALG"
   },
   "source": [
    "## 파이썬으로 아달린 구현하기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:56.582502Z",
     "iopub.status.busy": "2020-10-13T15:10:56.580804Z",
     "iopub.status.idle": "2020-10-13T15:10:56.585803Z",
     "shell.execute_reply": "2020-10-13T15:10:56.584255Z"
    },
    "id": "r_X9q8k_qALG"
   },
   "outputs": [],
   "source": [
    "class AdalineGD(object):\n",
    "    \"\"\"적응형 선형 뉴런 분류기\n",
    "\n",
    "    매개변수\n",
    "    ------------\n",
    "    eta : float\n",
    "      학습률 (0.0과 1.0 사이)\n",
    "    n_iter : int\n",
    "      훈련 데이터셋 반복 횟수\n",
    "    random_state : int\n",
    "      가중치 무작위 초기화를 위한 난수 생성기 시드\n",
    "\n",
    "    속성\n",
    "    -----------\n",
    "    w_ : 1d-array\n",
    "      학습된 가중치\n",
    "    cost_ : list\n",
    "      에포크마다 누적된 비용 함수의 제곱합\n",
    "\n",
    "    \"\"\"\n",
    "    def __init__(self, eta=0.01, n_iter=50, random_state=1):\n",
    "        self.eta = eta\n",
    "        self.n_iter = n_iter\n",
    "        self.random_state = random_state\n",
    "\n",
    "    def fit(self, X, y):\n",
    "        \"\"\"훈련 데이터 학습\n",
    "\n",
    "        매개변수\n",
    "        ----------\n",
    "        X : {array-like}, shape = [n_samples, n_features]\n",
    "          n_samples 개의 샘플과 n_features 개의 특성으로 이루어진 훈련 데이터\n",
    "        y : array-like, shape = [n_samples]\n",
    "          타깃값\n",
    "\n",
    "        반환값\n",
    "        -------\n",
    "        self : object\n",
    "\n",
    "        \"\"\"\n",
    "        rgen = np.random.RandomState(self.random_state)\n",
    "        self.w_ = rgen.normal(loc=0.0, scale=0.01, size=1 + X.shape[1])\n",
    "        self.cost_ = []\n",
    "\n",
    "        for i in range(self.n_iter):\n",
    "            net_input = self.net_input(X)\n",
    "            # 이 코드의 활성화 함수는 항등 함수(identity function)이기 때문에\n",
    "            # 아무런 효과가 없습니다.\n",
    "            # 이 대신 `output = self.net_input(X)`로 바로 쓸 수 있습니다.\n",
    "            # 이 활성화 함수는 개념적인 목적을 위해 만들었습니다.\n",
    "            # (잠시 후에 보게 될) 로지스틱 회귀의 경우 이 함수를 시그모이드 함수로 \n",
    "            # 바꾸어 로지스틱 회귀 분류기를 구현합니다.\n",
    "            output = self.activation(net_input)\n",
    "            errors = (y - output)\n",
    "            self.w_[1:] += self.eta * X.T.dot(errors)\n",
    "            self.w_[0] += self.eta * errors.sum()\n",
    "            cost = (errors**2).sum() / 2.0\n",
    "            self.cost_.append(cost)\n",
    "        return self\n",
    "\n",
    "    def net_input(self, X):\n",
    "        \"\"\"최종 입력 계산\"\"\"\n",
    "        return np.dot(X, self.w_[1:]) + self.w_[0]\n",
    "\n",
    "    def activation(self, X):\n",
    "        \"\"\"선형 활성화 계산\"\"\"\n",
    "        return X\n",
    "\n",
    "    def predict(self, X):\n",
    "        \"\"\"단위 계단 함수를 사용하여 클래스 레이블을 반환합니다\"\"\"\n",
    "        return np.where(self.activation(self.net_input(X)) >= 0.0, 1, -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:56.640703Z",
     "iopub.status.busy": "2020-10-13T15:10:56.615186Z",
     "iopub.status.idle": "2020-10-13T15:10:57.212231Z",
     "shell.execute_reply": "2020-10-13T15:10:57.213681Z"
    },
    "id": "duwyjN71qALG",
    "outputId": "0e540950-017a-4322-98c1-da2ccd6fe3ae"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots(nrows=1, ncols=2, figsize=(10, 4))\n",
    "\n",
    "ada1 = AdalineGD(n_iter=10, eta=0.01).fit(X, y)\n",
    "ax[0].plot(range(1, len(ada1.cost_) + 1), np.log10(ada1.cost_), marker='o')\n",
    "ax[0].set_xlabel('Epochs')\n",
    "ax[0].set_ylabel('log(Sum-squared-error)')\n",
    "ax[0].set_title('Adaline - Learning rate 0.01')\n",
    "\n",
    "ada2 = AdalineGD(n_iter=10, eta=0.0001).fit(X, y)\n",
    "ax[1].plot(range(1, len(ada2.cost_) + 1), ada2.cost_, marker='o')\n",
    "ax[1].set_xlabel('Epochs')\n",
    "ax[1].set_ylabel('Sum-squared-error')\n",
    "ax[1].set_title('Adaline - Learning rate 0.0001')\n",
    "\n",
    "# plt.savefig('images/02_11.png', dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "j09HnTnrqALH"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 261
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:57.222554Z",
     "iopub.status.busy": "2020-10-13T15:10:57.221110Z",
     "iopub.status.idle": "2020-10-13T15:10:57.263875Z",
     "shell.execute_reply": "2020-10-13T15:10:57.265089Z"
    },
    "id": "4w5q2QKqqALH",
    "outputId": "16c3e77a-9901-451c-8c83-043ed4d559f0"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<img src=\"https://git.io/JtIb4\" width=\"700\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 17,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(url='https://git.io/JtIb4', width=700) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "3i_pATdkqALH"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "6u0jOIB3qALH"
   },
   "source": [
    "## 특성 스케일을 조정하여 경사 하강법 결과 향상시키기"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 346
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:57.273097Z",
     "iopub.status.busy": "2020-10-13T15:10:57.271618Z",
     "iopub.status.idle": "2020-10-13T15:10:57.319363Z",
     "shell.execute_reply": "2020-10-13T15:10:57.320210Z"
    },
    "id": "PZ5nowNeqALI",
    "outputId": "06719022-db08-4ae0-8d3b-68bdfe55d6f5"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<img src=\"https://git.io/JtIbB\" width=\"700\"/>"
      ],
      "text/plain": [
       "<IPython.core.display.Image object>"
      ]
     },
     "execution_count": 18,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Image(url='https://git.io/JtIbB', width=700) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:57.332752Z",
     "iopub.status.busy": "2020-10-13T15:10:57.331310Z",
     "iopub.status.idle": "2020-10-13T15:10:57.336002Z",
     "shell.execute_reply": "2020-10-13T15:10:57.334757Z"
    },
    "id": "nEBXmUTkqALI"
   },
   "outputs": [],
   "source": [
    "# 특성을 표준화합니다.\n",
    "X_std = np.copy(X)\n",
    "X_std[:, 0] = (X[:, 0] - X[:, 0].mean()) / X[:, 0].std()\n",
    "X_std[:, 1] = (X[:, 1] - X[:, 1].mean()) / X[:, 1].std()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 577
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:57.352419Z",
     "iopub.status.busy": "2020-10-13T15:10:57.350540Z",
     "iopub.status.idle": "2020-10-13T15:10:58.151876Z",
     "shell.execute_reply": "2020-10-13T15:10:58.151183Z"
    },
    "id": "OGiKd5kvqALI",
    "outputId": "9d00703b-9e71-44ab-b1e9-c5a663efd5ac"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ada_gd = AdalineGD(n_iter=15, eta=0.01)\n",
    "ada_gd.fit(X_std, y)\n",
    "\n",
    "plot_decision_regions(X_std, y, classifier=ada_gd)\n",
    "plt.title('Adaline - Gradient Descent')\n",
    "plt.xlabel('sepal length [standardized]')\n",
    "plt.ylabel('petal length [standardized]')\n",
    "plt.legend(loc='upper left')\n",
    "plt.tight_layout()\n",
    "# plt.savefig('images/02_14_1.png', dpi=300)\n",
    "plt.show()\n",
    "\n",
    "plt.plot(range(1, len(ada_gd.cost_) + 1), ada_gd.cost_, marker='o')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Sum-squared-error')\n",
    "\n",
    "plt.tight_layout()\n",
    "# plt.savefig('images/02_14_2.png', dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "l9aKnN4fqALJ"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hj-m4ewrqALJ"
   },
   "source": [
    "## 대규모 머신 러닝과 확률적 경사 하강법"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:58.171932Z",
     "iopub.status.busy": "2020-10-13T15:10:58.171351Z",
     "iopub.status.idle": "2020-10-13T15:10:58.174440Z",
     "shell.execute_reply": "2020-10-13T15:10:58.173990Z"
    },
    "id": "Fmyd49AbqALJ"
   },
   "outputs": [],
   "source": [
    "class AdalineSGD(object):\n",
    "    \"\"\"ADAptive LInear NEuron 분류기\n",
    "\n",
    "    Parameters\n",
    "    ------------\n",
    "    eta : float\n",
    "      학습률 (0.0과 1.0 사이)\n",
    "    n_iter : int\n",
    "      훈련 데이터셋 반복 횟수\n",
    "    shuffle : bool (default: True)\n",
    "      True로 설정하면 같은 반복이 되지 않도록 에포크마다 훈련 데이터를 섞습니다\n",
    "    random_state : int\n",
    "      가중치 무작위 초기화를 위한 난수 생성기 시드\n",
    "\n",
    "    Attributes\n",
    "    -----------\n",
    "    w_ : 1d-array\n",
    "      학습된 가중치\n",
    "    cost_ : list\n",
    "      모든 훈련 샘플에 대해 에포크마다 누적된 평균 비용 함수의 제곱합\n",
    "\n",
    "    \"\"\"\n",
    "    def __init__(self, eta=0.01, n_iter=10, shuffle=True, random_state=None):\n",
    "        self.eta = eta\n",
    "        self.n_iter = n_iter\n",
    "        self.w_initialized = False\n",
    "        self.shuffle = shuffle\n",
    "        self.random_state = random_state\n",
    "        \n",
    "    def fit(self, X, y):\n",
    "        \"\"\"훈련 데이터 학습\n",
    "\n",
    "        Parameters\n",
    "        ----------\n",
    "        X : {array-like}, shape = [n_samples, n_features]\n",
    "          n_samples 개의 샘플과 n_features 개의 특성으로 이루어진 훈련 데이터\n",
    "        y : array-like, shape = [n_samples]\n",
    "          타깃 벡터\n",
    "\n",
    "        반환값\n",
    "        -------\n",
    "        self : object\n",
    "\n",
    "        \"\"\"\n",
    "        self._initialize_weights(X.shape[1])\n",
    "        self.cost_ = []\n",
    "        for i in range(self.n_iter):\n",
    "            if self.shuffle:\n",
    "                X, y = self._shuffle(X, y)\n",
    "            cost = []\n",
    "            for xi, target in zip(X, y):\n",
    "                cost.append(self._update_weights(xi, target))\n",
    "            avg_cost = sum(cost) / len(y)\n",
    "            self.cost_.append(avg_cost)\n",
    "        return self\n",
    "\n",
    "    def partial_fit(self, X, y):\n",
    "        \"\"\"가중치를 다시 초기화하지 않고 훈련 데이터를 학습합니다\"\"\"\n",
    "        if not self.w_initialized:\n",
    "            self._initialize_weights(X.shape[1])\n",
    "        if y.ravel().shape[0] > 1:\n",
    "            for xi, target in zip(X, y):\n",
    "                self._update_weights(xi, target)\n",
    "        else:\n",
    "            self._update_weights(X, y)\n",
    "        return self\n",
    "\n",
    "    def _shuffle(self, X, y):\n",
    "        \"\"\"훈련 데이터를 섞습니다\"\"\"\n",
    "        r = self.rgen.permutation(len(y))\n",
    "        return X[r], y[r]\n",
    "    \n",
    "    def _initialize_weights(self, m):\n",
    "        \"\"\"랜덤한 작은 수로 가중치를 초기화합니다\"\"\"\n",
    "        self.rgen = np.random.RandomState(self.random_state)\n",
    "        self.w_ = self.rgen.normal(loc=0.0, scale=0.01, size=1 + m)\n",
    "        self.w_initialized = True\n",
    "        \n",
    "    def _update_weights(self, xi, target):\n",
    "        \"\"\"아달린 학습 규칙을 적용하여 가중치를 업데이트합니다\"\"\"\n",
    "        output = self.activation(self.net_input(xi))\n",
    "        error = (target - output)\n",
    "        self.w_[1:] += self.eta * xi.dot(error)\n",
    "        self.w_[0] += self.eta * error\n",
    "        cost = 0.5 * error**2\n",
    "        return cost\n",
    "    \n",
    "    def net_input(self, X):\n",
    "        \"\"\"입력 계산\"\"\"\n",
    "        return np.dot(X, self.w_[1:]) + self.w_[0]\n",
    "\n",
    "    def activation(self, X):\n",
    "        \"\"\"선형 활성화 계산\"\"\"\n",
    "        return X\n",
    "\n",
    "    def predict(self, X):\n",
    "        \"\"\"단위 계단 함수를 사용하여 클래스 레이블을 반환합니다\"\"\"\n",
    "        return np.where(self.activation(self.net_input(X)) >= 0.0, 1, -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 577
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:58.182211Z",
     "iopub.status.busy": "2020-10-13T15:10:58.181685Z",
     "iopub.status.idle": "2020-10-13T15:10:59.021061Z",
     "shell.execute_reply": "2020-10-13T15:10:59.020293Z"
    },
    "id": "CL_zIPSEqALJ",
    "outputId": "663be36e-787e-41e4-a861-a0eda635e61d"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "ada_sgd = AdalineSGD(n_iter=15, eta=0.01, random_state=1)\n",
    "ada_sgd.fit(X_std, y)\n",
    "\n",
    "plot_decision_regions(X_std, y, classifier=ada_sgd)\n",
    "plt.title('Adaline - Stochastic Gradient Descent')\n",
    "plt.xlabel('sepal length [standardized]')\n",
    "plt.ylabel('petal length [standardized]')\n",
    "plt.legend(loc='upper left')\n",
    "\n",
    "plt.tight_layout()\n",
    "# plt.savefig('images/02_15_1.png', dpi=300)\n",
    "plt.show()\n",
    "\n",
    "plt.plot(range(1, len(ada_sgd.cost_) + 1), ada_sgd.cost_, marker='o')\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Average Cost')\n",
    "\n",
    "plt.tight_layout()\n",
    "# plt.savefig('images/02_15_2.png', dpi=300)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "execution": {
     "iopub.execute_input": "2020-10-13T15:10:59.028098Z",
     "iopub.status.busy": "2020-10-13T15:10:59.027571Z",
     "iopub.status.idle": "2020-10-13T15:10:59.032180Z",
     "shell.execute_reply": "2020-10-13T15:10:59.031725Z"
    },
    "id": "8MIHsvCsqALK",
    "outputId": "9309e527-eed2-4f8e-8aa4-9a6d6b9da4db"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<__main__.AdalineSGD at 0x7fb52010cb00>"
      ]
     },
     "execution_count": 23,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ada_sgd.partial_fit(X_std[0, :], y[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "H98ahObaqALK"
   },
   "source": [
    "<br>\n",
    "<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "HN8hlUaHqALL"
   },
   "source": [
    "# 요약"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "6zpIsTtMqALL"
   },
   "source": [
    "..."
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "colab": {
   "name": "ch02.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "TensorFlow 2.4 on Python 3.8 & CUDA 11.1",
   "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": "3.8.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}