贝叶斯判别

参考文献：

6 判别分析 | 多元统计分析示例
https://www.cnblogs.com/qizhou/p/13495598.html

一、问题描述

贝叶斯判别的本质是一类分类问题：基于若干采样样本，如何学习一个分类器对新样本数据进行分类并保证分类错误的概率最小。

假设

一共存在 n 个类别，分别为 $c_0,c_1,...,c_{n-1}$
每个类别都能够用相同的 p 个属性进行描述
对整体样本进行 m 次的采样，分别为 $x_0,x_1,...,x_{n-1}$
每个采样样本 $x_i$ 包含 p 个属性，即 $x_i = [f_0,f_1,...,f_{p-1}]$ ，并对应一个类别 $c_i$

贝叶斯假设某一个新样本 $x$ 属于每个类别 $c$ 为一个随机变量，服从概率分布 $p(c|x)$ ，那么只要分别计算出 $p(c_0|x),p(c_1|x),...,p(c_{n-1}|x)$ 的概率，那么概率最大的那个类别就是使分类错误概率最小的分类结果。

二、理论推导

后验概率 $p(c|x)$ 无法直接计算，因此根据贝叶斯公式，将后验概率转化为先验概率 $p(c)$ 与似然概率 $p(x|c)$ 乘积的形式.

$p(c|x)=\frac{p(c)p(x|c)}{p(x)}\sim p(c)p(x|c)$

先验概率 $p(c)$ 描述了每个类别在全体样本中的比例。在没有先验信息的条件下可以假定每个类别在样本中均匀分布，即 $p(c_i)=\frac{1}{n}$ ；也可以根据已有样本中不同类别出现的频率对类别的分布进行近似。

似然概率 $p(x|c)$ 描述了对应类别属性的分布，一般假定似然概率服从多元高斯分布，即

$p(x|c) = \frac{1}{\sqrt{(2\pi)^p\Sigma}}exp^{(-\frac{1}{2}(x-\mu)^T\Sigma^{-1}(x-\mu))}$

其中 $\mu$ 为每个类别对应属性的平均值， $\Sigma$ 为每个类别对应属性的协方差矩阵。

因此，后验概率

$p(c|x)\sim p(c)p(x|c)=p(c)\frac{1}{\sqrt{(2\pi)^p\Sigma}}exp^{(-\frac{1}{2}(x-\mu)^T\Sigma^{-1}(x-\mu))}$

进一步省略常数项，取对数得到

$p(c|x)\sim ln(p(c)) - \frac{1}{2}ln(| \Sigma|) - \frac{1}{2}(x-\mu)\Sigma^{-1}(x-\mu)^T$

分别得到不同类别情况下 $p(c|x)$ 的均值和方差，然后计算新样本在每个类别下的后验概率，后验概率最大的类别就是贝叶斯判别的类别。

三、后验概率均值和方差计算

后验概率的均值计算比较简单

$\mu_i = [\mu_i^0,\mu_i^1,...,\mu_i^{p-1}] = \frac{1}{m_i}\sum_{j=1}^{m_i}x_i^j$

其中 $\mu_i$ 是第 i 个类别所有特征的均值，为一个 p 维向量，分别对应每个特性的均值； $m_i$ 表示第 i 类别一共采样了 $m_i$ 个样本；x_i^j 表示第 j 个属于第 i 个类别的样本。

后验概率的协方差 $\Sigma=\frac{1}{m_i-1}B^TB$ ，其中

$B=\begin{bmatrix} x_i^0-\mu_i \\ x_i^1-\mu_i \\ ... \\ x_i^{m_{i-1}}-\mu_i \\ \end{bmatrix}$ 为一个 $m_i \times p$ 维的矩阵。

四、代码实现

Python 实现

import numpy as nptmp = np.loadtxt("iris.csv", dtype=str, delimiter=",")
data = tmp[1:,1:-1].astype(float)
label = np.array([['"setosa"', '"versicolor"', '"virginica"'].index(t) for t in tmp[1:, -1]])data_0 = data[label == 0, :]
m_data_0 = np.mean(data_0, axis=0)
sigma_0 = np.matmul((data_0 - m_data_0).T, (data_0 - m_data_0)) / (len(data_0) - 1)data_1 = data[label == 1, :]
m_data_1 = np.mean(data_1, axis=0)
sigma_1 = np.matmul((data_1 - m_data_1).T, (data_1 - m_data_1)) / (len(data_1) - 1)data_2 = data[label == 2, :]
m_data_2 = np.mean(data_2, axis=0)
sigma_2 = np.matmul((data_2 - m_data_2).T, (data_2 - m_data_2)) / (len(data_2) - 1)d = data[-1]
p_0 = np.log(50 / 150) - 0.5 * np.log(np.linalg.det(sigma_0)) - 0.5 * (d - m_data_0) @ np.linalg.inv(sigma_0) @ (d - m_data_0).T
p_1 = np.log(50 / 150) - 0.5 * np.log(np.linalg.det(sigma_1)) - 0.5 * (d - m_data_1) @ np.linalg.inv(sigma_1) @ (d - m_data_1).T
p_2 = np.log(50 / 150) - 0.5 * np.log(np.linalg.det(sigma_2)) - 0.5 * (d - m_data_2) @ np.linalg.inv(sigma_2) @ (d - m_data_2).T

matlab 实现

clc;
clear;tmp = importdata('iris.csv', ',', 1);
data = tmp.data(:, 1:end-1);
label = zeros(1,150);
label(51:100) = 1;
label(101:end) = 2;data_0 = data(label == 0, :);
m_data_0 = mean(data_0, dim=1);
sigma_0 = cov(data_0);data_1 = data(label == 1, :);
m_data_1 = mean(data_1, dim=1);
sigma_1 = cov(data_1);data_2 = data(label == 2, :);
m_data_2 = mean(data_2, dim=1);
sigma_2 = cov(data_2);d = data(end, :);
p_0 = log(50 / 150) - 0.5 * log(det(sigma_0)) - 0.5 * (d - m_data_0) * inv(sigma_0) * (d - m_data_0)'
p_1 = log(50 / 150) - 0.5 * log(det(sigma_1)) - 0.5 * (d - m_data_1) * inv(sigma_1) * (d - m_data_1)'
p_2 = log(50 / 150) - 0.5 * log(det(sigma_2)) - 0.5 * (d - m_data_2) * inv(sigma_2) * (d - m_data_2)'

测试数据 iris.csv

"Sepal.Length" "Sepal.Width" "Petal.Length" "Petal.Width" "Species"
"1" 5.1 3.5 1.4 0.2 "setosa"
"2" 4.9 3 1.4 0.2 "setosa"
"3" 4.7 3.2 1.3 0.2 "setosa"
"4" 4.6 3.1 1.5 0.2 "setosa"
"5" 5 3.6 1.4 0.2 "setosa"
"6" 5.4 3.9 1.7 0.4 "setosa"
"7" 4.6 3.4 1.4 0.3 "setosa"
"8" 5 3.4 1.5 0.2 "setosa"
"9" 4.4 2.9 1.4 0.2 "setosa"
"10" 4.9 3.1 1.5 0.1 "setosa"
"11" 5.4 3.7 1.5 0.2 "setosa"
"12" 4.8 3.4 1.6 0.2 "setosa"
"13" 4.8 3 1.4 0.1 "setosa"
"14" 4.3 3 1.1 0.1 "setosa"
"15" 5.8 4 1.2 0.2 "setosa"
"16" 5.7 4.4 1.5 0.4 "setosa"
"17" 5.4 3.9 1.3 0.4 "setosa"
"18" 5.1 3.5 1.4 0.3 "setosa"
"19" 5.7 3.8 1.7 0.3 "setosa"
"20" 5.1 3.8 1.5 0.3 "setosa"
"21" 5.4 3.4 1.7 0.2 "setosa"
"22" 5.1 3.7 1.5 0.4 "setosa"
"23" 4.6 3.6 1 0.2 "setosa"
"24" 5.1 3.3 1.7 0.5 "setosa"
"25" 4.8 3.4 1.9 0.2 "setosa"
"26" 5 3 1.6 0.2 "setosa"
"27" 5 3.4 1.6 0.4 "setosa"
"28" 5.2 3.5 1.5 0.2 "setosa"
"29" 5.2 3.4 1.4 0.2 "setosa"
"30" 4.7 3.2 1.6 0.2 "setosa"
"31" 4.8 3.1 1.6 0.2 "setosa"
"32" 5.4 3.4 1.5 0.4 "setosa"
"33" 5.2 4.1 1.5 0.1 "setosa"
"34" 5.5 4.2 1.4 0.2 "setosa"
"35" 4.9 3.1 1.5 0.2 "setosa"
"36" 5 3.2 1.2 0.2 "setosa"
"37" 5.5 3.5 1.3 0.2 "setosa"
"38" 4.9 3.6 1.4 0.1 "setosa"
"39" 4.4 3 1.3 0.2 "setosa"
"40" 5.1 3.4 1.5 0.2 "setosa"
"41" 5 3.5 1.3 0.3 "setosa"
"42" 4.5 2.3 1.3 0.3 "setosa"
"43" 4.4 3.2 1.3 0.2 "setosa"
"44" 5 3.5 1.6 0.6 "setosa"
"45" 5.1 3.8 1.9 0.4 "setosa"
"46" 4.8 3 1.4 0.3 "setosa"
"47" 5.1 3.8 1.6 0.2 "setosa"
"48" 4.6 3.2 1.4 0.2 "setosa"
"49" 5.3 3.7 1.5 0.2 "setosa"
"50" 5 3.3 1.4 0.2 "setosa"
"51" 7 3.2 4.7 1.4 "versicolor"
"52" 6.4 3.2 4.5 1.5 "versicolor"
"53" 6.9 3.1 4.9 1.5 "versicolor"
"54" 5.5 2.3 4 1.3 "versicolor"
"55" 6.5 2.8 4.6 1.5 "versicolor"
"56" 5.7 2.8 4.5 1.3 "versicolor"
"57" 6.3 3.3 4.7 1.6 "versicolor"
"58" 4.9 2.4 3.3 1 "versicolor"
"59" 6.6 2.9 4.6 1.3 "versicolor"
"60" 5.2 2.7 3.9 1.4 "versicolor"
"61" 5 2 3.5 1 "versicolor"
"62" 5.9 3 4.2 1.5 "versicolor"
"63" 6 2.2 4 1 "versicolor"
"64" 6.1 2.9 4.7 1.4 "versicolor"
"65" 5.6 2.9 3.6 1.3 "versicolor"
"66" 6.7 3.1 4.4 1.4 "versicolor"
"67" 5.6 3 4.5 1.5 "versicolor"
"68" 5.8 2.7 4.1 1 "versicolor"
"69" 6.2 2.2 4.5 1.5 "versicolor"
"70" 5.6 2.5 3.9 1.1 "versicolor"
"71" 5.9 3.2 4.8 1.8 "versicolor"
"72" 6.1 2.8 4 1.3 "versicolor"
"73" 6.3 2.5 4.9 1.5 "versicolor"
"74" 6.1 2.8 4.7 1.2 "versicolor"
"75" 6.4 2.9 4.3 1.3 "versicolor"
"76" 6.6 3 4.4 1.4 "versicolor"
"77" 6.8 2.8 4.8 1.4 "versicolor"
"78" 6.7 3 5 1.7 "versicolor"
"79" 6 2.9 4.5 1.5 "versicolor"
"80" 5.7 2.6 3.5 1 "versicolor"
"81" 5.5 2.4 3.8 1.1 "versicolor"
"82" 5.5 2.4 3.7 1 "versicolor"
"83" 5.8 2.7 3.9 1.2 "versicolor"
"84" 6 2.7 5.1 1.6 "versicolor"
"85" 5.4 3 4.5 1.5 "versicolor"
"86" 6 3.4 4.5 1.6 "versicolor"
"87" 6.7 3.1 4.7 1.5 "versicolor"
"88" 6.3 2.3 4.4 1.3 "versicolor"
"89" 5.6 3 4.1 1.3 "versicolor"
"90" 5.5 2.5 4 1.3 "versicolor"
"91" 5.5 2.6 4.4 1.2 "versicolor"
"92" 6.1 3 4.6 1.4 "versicolor"
"93" 5.8 2.6 4 1.2 "versicolor"
"94" 5 2.3 3.3 1 "versicolor"
"95" 5.6 2.7 4.2 1.3 "versicolor"
"96" 5.7 3 4.2 1.2 "versicolor"
"97" 5.7 2.9 4.2 1.3 "versicolor"
"98" 6.2 2.9 4.3 1.3 "versicolor"
"99" 5.1 2.5 3 1.1 "versicolor"
"100" 5.7 2.8 4.1 1.3 "versicolor"
"101" 6.3 3.3 6 2.5 "virginica"
"102" 5.8 2.7 5.1 1.9 "virginica"
"103" 7.1 3 5.9 2.1 "virginica"
"104" 6.3 2.9 5.6 1.8 "virginica"
"105" 6.5 3 5.8 2.2 "virginica"
"106" 7.6 3 6.6 2.1 "virginica"
"107" 4.9 2.5 4.5 1.7 "virginica"
"108" 7.3 2.9 6.3 1.8 "virginica"
"109" 6.7 2.5 5.8 1.8 "virginica"
"110" 7.2 3.6 6.1 2.5 "virginica"
"111" 6.5 3.2 5.1 2 "virginica"
"112" 6.4 2.7 5.3 1.9 "virginica"
"113" 6.8 3 5.5 2.1 "virginica"
"114" 5.7 2.5 5 2 "virginica"
"115" 5.8 2.8 5.1 2.4 "virginica"
"116" 6.4 3.2 5.3 2.3 "virginica"
"117" 6.5 3 5.5 1.8 "virginica"
"118" 7.7 3.8 6.7 2.2 "virginica"
"119" 7.7 2.6 6.9 2.3 "virginica"
"120" 6 2.2 5 1.5 "virginica"
"121" 6.9 3.2 5.7 2.3 "virginica"
"122" 5.6 2.8 4.9 2 "virginica"
"123" 7.7 2.8 6.7 2 "virginica"
"124" 6.3 2.7 4.9 1.8 "virginica"
"125" 6.7 3.3 5.7 2.1 "virginica"
"126" 7.2 3.2 6 1.8 "virginica"
"127" 6.2 2.8 4.8 1.8 "virginica"
"128" 6.1 3 4.9 1.8 "virginica"
"129" 6.4 2.8 5.6 2.1 "virginica"
"130" 7.2 3 5.8 1.6 "virginica"
"131" 7.4 2.8 6.1 1.9 "virginica"
"132" 7.9 3.8 6.4 2 "virginica"
"133" 6.4 2.8 5.6 2.2 "virginica"
"134" 6.3 2.8 5.1 1.5 "virginica"
"135" 6.1 2.6 5.6 1.4 "virginica"
"136" 7.7 3 6.1 2.3 "virginica"
"137" 6.3 3.4 5.6 2.4 "virginica"
"138" 6.4 3.1 5.5 1.8 "virginica"
"139" 6 3 4.8 1.8 "virginica"
"140" 6.9 3.1 5.4 2.1 "virginica"
"141" 6.7 3.1 5.6 2.4 "virginica"
"142" 6.9 3.1 5.1 2.3 "virginica"
"143" 5.8 2.7 5.1 1.9 "virginica"
"144" 6.8 3.2 5.9 2.3 "virginica"
"145" 6.7 3.3 5.7 2.5 "virginica"
"146" 6.7 3 5.2 2.3 "virginica"
"147" 6.3 2.5 5 1.9 "virginica"
"148" 6.5 3 5.2 2 "virginica"
"149" 6.2 3.4 5.4 2.3 "virginica"
"150" 5.9 3 5.1 1.8 "virginica"