2_Naive_Bayes_Classification.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zOgE_hxI--32"
      },
      "source": [
        "# Naive Bayes Classification\n",
        " \n",
        "\n",
        "Naive Bayes Classification 은 모든 특성 값이 서로 독립임을 가정하여, 특성들 사이에서의 발생할 수 있는 연관성을 배제하고 각각의 특성이 분류할 대상에 대해 독립적으로 기여할 수 있도록 한다. 베이즈 정리와 조건부 확률을 이용하여 다음과 같이 사후 확률 (Posterior)에 대하여 사전 확률 (Prior) 과 우도 (Likelihood) 로 표현할 수 있다. \n",
        "\n",
        "* \\\\( \\hat{y} = \\text{argmax}_y p(y | X) \\\\)  \n",
        "* \\\\( p(y|X) = \\frac{p( \\mathbf{X} | y) p(y)}{p(\\mathbf{X})} \\\\)  \n",
        "    * 사전 확률 (Prior Probability) : \\\\( p(y) \\\\)\n",
        "    * 우도 (Likelihood Probability) : \\\\( p(X | y) =  \\prod_i p(x_i | y) p(y) \\;  \\\\)   \n",
        "    * 사후 확률 (Posterior Probability) : \\\\( p(y | X)  \\\\)\n",
        "    \n",
        "(Naive Bayes Classifier : https://en.wikipedia.org/wiki/Naive_Bayes_classifier)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BsNI-YWH--34"
      },
      "source": [
        "#### < 예제 > Naive Bayes Classification 을 이용한 Spam filter\n",
        "\n",
        "* 조건\n",
        "\n",
        "    * 스팸 필터링을 위한 Naive Bayes model 의 독립 변수 (V)\n",
        "        * V = ”secret”, ”offer”, ”low”, ”price”, ”valued”, ”customer”, ”today”, ”dollar”, ”million”, ”sports”, ”is”, ”for”, ”play”, ”healthy”, ”pizza”\n",
        "\n",
        "    * 주어진 학습 데이터 (스팸 메시지 3개, 정상 메시지 4개)\n",
        "        * 스팸 메시지 = ”million dollar offer”, ”secret offer today”, ”secret is secret”\n",
        "        * 정상 메시지 = ”low price for valued customer”, ”play secret sports today”, ”sports is healthy”, ”low price pizza”\n",
        "\n",
        "\n",
        "*  Naive Bayes model 과 학습 데이터로 제공된 메시지를 이용하여 각각의 확률을 계산\n",
        "    * P(secret|spam) := Nd(secret,spam) / Nd(spam) = 2/3\n",
        "    * P(spam) = ?\n",
        "    * P(sports|non-spam) = ? \n",
        "    * P(dollar|spam) = ?\n",
        "    * P(“sports is healthy”|spam) = ? \n",
        "    * P(“sports is healthy”|non-spam) = ?\n",
        "    * P(spam|“sports is healthy”) = ? \n",
        "    * P(non-spam|“sports is healthy”) = ? \n",
        "\n",
        "\n",
        "* 학습된 Naive Bayes model 을 이용하여 스팸 메시지와 정상 메시지를 구분하는 코드 구현\n",
        "    * \\\\( p(y|X) = \\frac{p( \\mathbf{X} | y) p(y)}{p(\\mathbf{X})} \\\\)\n",
        "    * matplotlib 라이브러리를 이용하여 visualization\n",
        "      \n",
        " \n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "W01g_zLH--36"
      },
      "outputs": [],
      "source": [
        "import numpy as np"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "id": "g7pbKAYS--36"
      },
      "outputs": [],
      "source": [
        "# Naive bayes model\n",
        "\n",
        "V = [\"secret\", \"offer\", \"low\", \"price\", \"valued\", \"customer\", \"today\", \"dollar\", \"million\", \"sports\", \"is\", \"for\", \"play\", \"healthy\", \"pizza\"]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "id": "JFdXnzba--37"
      },
      "outputs": [],
      "source": [
        "# 단어 별 인덱스 추가\n",
        "\n",
        "v_idx_mapping = {}\n",
        "for i, word in enumerate(V):\n",
        "    v_idx_mapping[word] = i\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "scrolled": true,
        "id": "vb7qqaIW--37",
        "outputId": "9b5b655d-9ae8-4d0f-e10b-c5575bc8c058",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'secret': 0,\n",
              " 'offer': 1,\n",
              " 'low': 2,\n",
              " 'price': 3,\n",
              " 'valued': 4,\n",
              " 'customer': 5,\n",
              " 'today': 6,\n",
              " 'dollar': 7,\n",
              " 'million': 8,\n",
              " 'sports': 9,\n",
              " 'is': 10,\n",
              " 'for': 11,\n",
              " 'play': 12,\n",
              " 'healthy': 13,\n",
              " 'pizza': 14}"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ],
      "source": [
        "v_idx_mapping"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "hik6e7UE--37"
      },
      "outputs": [],
      "source": [
        "# 주어진 학습 데이터 \n",
        "# 스팸 메시지 : spam_msgs\n",
        "# 정상 메시지 : ham_msgs \n",
        "\n",
        "spam_msgs = [\"million dollar offer\", \"secret offer today\", \"secret is secret\"]\n",
        "ham_msgs = [\"low price for valued customer offer\", \"play secret sports today\", \"sports is healthy\", \"low price pizza\"]\n",
        "\n",
        "# ' '.join(spam_msgs)\n",
        "# TEXT.split(' ')\n",
        "\n",
        "# msgs = spam_msgs + ham_msgs\n",
        "# words = ' '.join(msgs).split(' ')\n",
        "# words = list(set(words))\n",
        "# print(len(words))\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "UELENA_g--38",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fe378cfb-65c5-456b-e541-a3176129cffe"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[{'million', 'dollar', 'offer'}, {'offer', 'secret', 'today'}, {'secret', 'is'}]\n",
            "[{'low', 'valued', 'for', 'offer', 'customer', 'price'}, {'sports', 'secret', 'play', 'today'}, {'healthy', 'sports', 'is'}, {'pizza', 'price', 'low'}]\n"
          ]
        }
      ],
      "source": [
        "table = np.zeros([2, len(V)])\n",
        "\n",
        "# 주어진 메시지를 단어 단위로 구분\n",
        "# 중복 단어 제거 \n",
        "\n",
        "spam_list = [ set(spam_msg.split(' ')) for spam_msg in spam_msgs ]\n",
        "ham_list = [ set(ham_msg.split(' ')) for ham_msg in ham_msgs ]\n",
        "print(spam_list)\n",
        "print(ham_list)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "V"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PHCe8D1PC_kq",
        "outputId": "62316e53-cc11-494b-87dc-61383d0334e3"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['secret',\n",
              " 'offer',\n",
              " 'low',\n",
              " 'price',\n",
              " 'valued',\n",
              " 'customer',\n",
              " 'today',\n",
              " 'dollar',\n",
              " 'million',\n",
              " 'sports',\n",
              " 'is',\n",
              " 'for',\n",
              " 'play',\n",
              " 'healthy',\n",
              " 'pizza']"
            ]
          },
          "metadata": {},
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "Uz4nbXEY--38"
      },
      "outputs": [],
      "source": [
        "# 스팸 메시지와 정상 메시지를 구성하는 단어의 수를 계산하여 Table 에 입력\n",
        "# 스팸 : [0, :] / 정상 : [1, :]\n",
        "\n",
        "for i, word in enumerate(V):\n",
        "    for spam_msg in spam_list:\n",
        "        table[0][v_idx_mapping[word]] += word in spam_msg\n",
        "    \n",
        "    for ham_msg in ham_list:\n",
        "        table[1][i] += word in ham_msg"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "4llhOnr6--38",
        "outputId": "19a5d689-cb17-405e-b083-af56cedbc830",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[2., 2., 0., 0., 0., 0., 1., 1., 1., 0., 1., 0., 0., 0., 0.],\n",
              "       [1., 1., 2., 2., 1., 1., 1., 0., 0., 2., 1., 1., 1., 1., 1.]])"
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ],
      "source": [
        "# y label : 스팸 메시지, 정상 메시지\n",
        "# X data : Naive Bayes model (V)\n",
        "# tabel : 학습 데이터로부터 계산한 스팸 메시지와 정상 메시지의 단어 수\n",
        "\n",
        "table"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "MAaATd8A--39"
      },
      "outputs": [],
      "source": [
        "# num_spam : 스팸 메시지 수\n",
        "# num_ham : 정상 메시지 수\n",
        "# num_total : 총 메시지 수\n",
        "\n",
        "num_spam = len(spam_msgs)\n",
        "num_ham = len(ham_msgs)\n",
        "num_total = num_ham + num_spam\n",
        "\n",
        "\n",
        "# 사전 확률 (prior) : P(spam), P(ham) \n",
        "\n",
        "p_spam = num_spam / num_total # P(Spam) = 3/7\n",
        "p_ham = num_ham / num_total # P(Non-Spam) = 4/7"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "XokMA0Pn--39",
        "outputId": "cc400761-2b96-4dfa-c83c-479d15ae51a9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "P(Xi | y) : \n",
            " [[0.66666667 0.66666667 0.         0.         0.         0.\n",
            "  0.33333333 0.33333333 0.33333333 0.         0.33333333 0.\n",
            "  0.         0.         0.        ]\n",
            " [0.25       0.25       0.5        0.5        0.25       0.25\n",
            "  0.25       0.         0.         0.5        0.25       0.25\n",
            "  0.25       0.25       0.25      ]]\n",
            "P(y) : \n",
            " [0.42857143 0.57142857]\n"
          ]
        }
      ],
      "source": [
        "# p_xi_y : 그 메시지가 spam 이라는걸 알때, xi(단어)가 등장할 확률\n",
        "# p_y : 전체 메시지 수에서 spam 일 확률, ham 일 확률\n",
        "p_xi_y = table / np.array([num_spam, num_ham]).reshape(2, 1)\n",
        "p_y = np.array([num_spam, num_ham]) / num_total\n",
        "\n",
        "print('\\nP(Xi | y) : \\n', p_xi_y)\n",
        "print('P(y) : \\n', p_y)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "G2TPU6Sb--39",
        "outputId": "6f8ae418-ce25-4111-bccb-41c766f406ee",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "P(secret|spam):  0.6666666666666666\n",
            "P(spam):  0.42857142857142855\n",
            "P(sports|non-spam):  0.5\n",
            "P(dollar|spam):  0.3333333333333333\n"
          ]
        }
      ],
      "source": [
        "# P(secret|spam) = Nd(secret, spam) / Nd(spam) = 2/3\n",
        "print('P(secret|spam): ', table[0][v_idx_mapping['secret']]/num_spam)\n",
        "print('P(spam): ', p_spam)\n",
        "\n",
        "p_xi_ham = table[1][v_idx_mapping['sports']]/num_ham\n",
        "p_xi_spam = table[0][v_idx_mapping['dollar']]/num_spam\n",
        "print('P(sports|non-spam): ', p_xi_ham)\n",
        "print('P(dollar|spam): ', p_xi_spam)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "1T4tLd4u--3-",
        "outputId": "fa789928-cf8a-4833-c0e5-8fd91e4e8c21",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "P(\"sports is healthy\"|spam):  0.0\n",
            "P(\"sports is healthy\"|non-spam):  0.0007821321487426758\n"
          ]
        }
      ],
      "source": [
        "p_x_spam = table[0][v_idx_mapping['sports']]/num_spam \\\n",
        "        * table[0][v_idx_mapping['is']]/num_spam \\\n",
        "        * table[0][v_idx_mapping['healthy']]/num_spam \\\n",
        "        * (1 - table[0][v_idx_mapping['secret']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['offer']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['low']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['price']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['valued']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['customer']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['today']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['dollar']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['million']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['for']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['play']]/num_spam) \\\n",
        "        * (1 - table[0][v_idx_mapping['pizza']]/num_spam)\n",
        "\n",
        "p_x_ham = table[1][v_idx_mapping['sports']]/num_ham \\\n",
        "        * table[1][v_idx_mapping['is']]/num_ham \\\n",
        "        * table[1][v_idx_mapping['healthy']]/num_ham \\\n",
        "        * (1 - table[1][v_idx_mapping['secret']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['offer']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['low']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['price']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['valued']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['customer']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['today']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['dollar']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['million']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['for']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['play']]/num_ham) \\\n",
        "        * (1 - table[1][v_idx_mapping['pizza']]/num_ham)\n",
        "\n",
        "print('P(\"sports is healthy\"|spam): ', p_x_spam)\n",
        "print('P(\"sports is healthy\"|non-spam): ', p_x_ham)\n"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# 위에처럼 일일이 다 작성하기는 너무 길다\n",
        "# email 을 binary vector 로 변환하는 함수\n",
        "\n",
        "def str2vec(email):\n",
        "  output = np.zeros(15, dtype=int)\n",
        "  for word in email.split():\n",
        "    output[v_idx_mapping[word]] = 1\n",
        "  return output\n"
      ],
      "metadata": {
        "id": "dc4VczpxuUCQ"
      },
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "str2vec(\"sports is healthy\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OI-VNCPPxjDC",
        "outputId": "1c16d4df-969d-4fc1-9397-32fa50874631"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0])"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "bin_email = str2vec(\"sports is healthy\")\n",
        "p_x_spam = np.prod(p_xi_y[0] * bin_email + (1 - p_xi_y[0]) * (1 - bin_email))\n",
        "p_x_ham = np.prod(p_xi_y[1] * bin_email + (1 - p_xi_y[1]) * (1 - bin_email))\n",
        "print('P(\"sports is healthy\"|spam): ', p_x_spam)\n",
        "print('P(\"sports is healthy\"|non-spam): ', p_x_ham)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wj5_ceVYvMuF",
        "outputId": "02a63e1d-61d0-4171-ecee-40d88614a727"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "P(\"sports is healthy\"|spam):  0.0\n",
            "P(\"sports is healthy\"|non-spam):  0.0007821321487426758\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "86zcElAB--3-",
        "outputId": "392671cf-a6c5-4f97-b010-6eca5454ec5c",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "P(spam|\"sports is healthy\"):  0.0\n",
            "P(non-spam|\"sports is healthy\"):  1.0\n"
          ]
        }
      ],
      "source": [
        "p_spam_x = (p_x_spam * p_spam) / ( (p_x_spam * p_spam) + (p_x_ham * p_ham) )\n",
        "p_ham_x = (p_x_ham * p_ham)  / ( (p_x_spam * p_spam) + (p_x_ham * p_ham) )\n",
        "\n",
        "print('P(spam|\"sports is healthy\"): ', p_spam_x)\n",
        "print('P(non-spam|\"sports is healthy\"): ', p_ham_x)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "id": "9Z5ja9u2--3_"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "* 조건\n",
        "\n",
        "    * 스팸 필터링을 위한 Naive Bayes model 의 독립 변수 (V)\n",
        "        * V = ”secret”, ”offer”, ”low”, ”price”, ”valued”, ”customer”, ”today”, ”dollar”, ”million”, ”sports”, ”is”, ”for”, ”play”, ”healthy”, ”pizza”\n",
        "\n",
        "    * 주어진 학습 데이터 (스팸 메시지 3개, 정상 메시지 4개)\n",
        "        * 스팸 메시지 = ”million dollar offer”, ”secret offer today”, ”secret is secret”\n",
        "        * 정상 메시지 = ”low price for valued customer”, ”play secret sports today”, ”sports is healthy”, ”low price pizza”\n",
        "\n",
        "\n",
        "* Laplace smoothing\n",
        "  * training data 에 없던 값들이 들어오거나 이상값들이 들어오면 그에 대한 likelihood가 0이 되어서 정상적으로 분류가 되지 않음. 이런 경우를 막기 위해 laplace smoothing 방법 사용. 상수 값을 더해 확률 값이 0이 나오지 않게 하여 조금이라도 해당 메일이 스팸 메일이라고 분류된 가능성을 증가시킴.\n",
        "  * $$p(x_i|y=spam) = \\frac{\\text{i번째 word가 등장한 메시지의 개수}+\\alpha}{\\text{spam 메시지 개수} + (\\alpha*K)}$$"
      ],
      "metadata": {
        "id": "8LFt7QMLE9Hq"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "lsGb3bYgS3Ke"
      },
      "outputs": [],
      "source": [
        "def naive_bayes_classifier(test_msg, table, p_y, laplace_smoothing = False):\n",
        "\n",
        "    print('input message : ', test_msg)\n",
        "    \n",
        "    # 주어진 메시지를 단어 단위로 구분하고 카운트 (중복 제거)\n",
        "    test_list = set(test_msg.split(' '))\n",
        "    table_test = np.zeros(len(V))\n",
        "\n",
        "    # 입력 메시지에서 사용된 단어 0 또는 1로 표시\n",
        "    table_test = str2vec(test_msg)\n",
        "\n",
        "    print(table_test)\n",
        "    # likelihood\n",
        "    alpha = 0.1\n",
        "    K = 1\n",
        "    if laplace_smoothing == True:\n",
        "      p_xi_y = (table+alpha) / np.array([num_spam+(alpha*K), num_ham+(alpha*K)]).reshape(2, 1)\n",
        "    else: \n",
        "      p_xi_y = table / np.array([num_spam, num_ham]).reshape(2, 1)\n",
        "\n",
        "    tmp = p_xi_y * table_test + (1 - p_xi_y) * (1 - table_test)\n",
        "\n",
        "    p_x_y = tmp.prod(1)\n",
        "    print('P(X|Spam): ', p_x_y[0])\n",
        "    print('P(X|Ham): ', p_x_y[1])\n",
        "    \n",
        "    # Posterior = likelihood x prior\n",
        "    p_y_x = p_x_y * p_y\n",
        "    p_y_x /= p_y_x.sum()\n",
        "    print('P(Spam|X): ', p_y_x[0])\n",
        "    print('P(Ham|X): ', p_y_x[1])\n",
        "    \n",
        "    if p_y_x[0] == 0.5:\n",
        "        print(\"cannot decide\")\n",
        "    else:\n",
        "        if list(p_y_x).index(max(p_y_x)) == 0:\n",
        "            print(\"Spam message!!\")\n",
        "        else:\n",
        "            print(\"Ham message!!\")\n",
        "    \n",
        "    # visualize\n",
        "    fig = plt.figure()\n",
        "    ax = fig.add_axes([0, 0, 1, 1])\n",
        "    ax.bar(list([\"spam\", \"ham\"]), p_y_x)\n",
        "    plt.title('Naive Bayes')\n",
        "    plt.show()\n",
        "    plt.close()\n",
        "    \n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "id": "llvEHl4T--4A",
        "outputId": "213a755c-22ba-46b1-83db-e9dd9bf9bb7b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 505
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "input message :  secret is secret\n",
            "[1 0 0 0 0 0 0 0 0 0 1 0 0 0 0]\n",
            "P(X|Spam):  0.021947873799725657\n",
            "P(X|Ham):  0.0007821321487426758\n",
            "[0.42857143 0.57142857]\n",
            "[0.02194787 0.00078213]\n",
            "P(Spam|X):  0.9546406977999647\n",
            "P(Ham|X):  0.04535930220003524\n",
            "Spam message!!\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# 스팸 메시지 확인\n",
        "naive_bayes_classifier(\"secret is secret\", table, p_y)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "id": "0XAH-AKI--4A",
        "outputId": "5e41979c-b4ed-439a-820d-8737dbf1748d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 505
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "input message :  sports is healthy\n",
            "[0 0 0 0 0 0 0 0 0 1 1 0 0 1 0]\n",
            "P(X|Spam):  0.0\n",
            "P(X|Ham):  0.0007821321487426758\n",
            "[0.42857143 0.57142857]\n",
            "[0.         0.00078213]\n",
            "P(Spam|X):  0.0\n",
            "P(Ham|X):  1.0\n",
            "Ham message!!\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# 정상 메시지 확인\n",
        "naive_bayes_classifier(\"sports is healthy\", table, p_y)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### P(X | Spam) = 0...\n",
        "### Need Laplace Smoothing!\n"
      ],
      "metadata": {
        "id": "jKcanH6CGvsx"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "id": "WeJvMMhZ--4B",
        "outputId": "ffdcf2c7-ea53-45ea-edec-0f731a4089e0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 505
        }
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "input message :  sports is healthy\n",
            "[0 0 0 0 0 0 0 0 0 1 1 0 0 1 0]\n",
            "P(X|Spam):  8.201740942286845e-06\n",
            "P(X|Ham):  0.0006861148608509721\n",
            "[0.42857143 0.57142857]\n",
            "[8.20174094e-06 6.86114861e-04]\n",
            "P(Spam|X):  0.008885752336132029\n",
            "P(Ham|X):  0.9911142476638679\n",
            "Ham message!!\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# practical naive bayes : laplace smoothing 적용\n",
        "naive_bayes_classifier(\"sports is healthy\", table, p_y, laplace_smoothing = True)"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Pymwy0E9BF8G"
      },
      "execution_count": 23,
      "outputs": []
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "aita",
      "language": "python",
      "name": "aita"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.12"
    },
    "colab": {
      "provenance": [],
      "collapsed_sections": []
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}