<Machine Learning in Action >之二 朴素贝叶斯 C#实现文章分类
def trainNB0(trainMatrix,trainCategory):
numTrainDocs = len(trainMatrix)
numWords = len(trainMatrix[0])
pAbusive = sum(trainCategory)/float(numTrainDocs)
p0Num = ones(numWords); p1Num = ones(numWords) #change to ones()
p0Denom = 2.0; p1Denom = 2.0 #change to 2.0
for i in range(numTrainDocs):
if trainCategory[i] == 1:
p1Num += trainMatrix[i]
p1Denom += sum(trainMatrix[i])
else:
p0Num += trainMatrix[i]
p0Denom += sum(trainMatrix[i])
p1Vect = log(p1Num/p1Denom) #change to log()
p0Vect = log(p0Num/p0Denom) #change to log()
return p0Vect,p1Vect,pAbusivedef classifyNB(vec2Classify, p0Vec, p1Vec, pClass1):
p1 = sum(vec2Classify * p1Vec) + log(pClass1) #element-wise mult *提示一
p0 = sum(vec2Classify * p0Vec) + log(1.0 - pClass1)
if p1 > p0:
return 1
else:
return 0
*提示一
p(Ci|w)=p(w|Ci)p(Ci)/p(w) 对乘积取自然对数 ln(p(w|Ci)p(Ci))=ln(p(w|Ci))+ln(p(Ci))
在以下样例中。由于每一个分类在样本中的比例都一样的，这样不用再加上log(p(Ci))也不会影响最后的分类效果
用C#随便做个样例，实现文章类型的分类 随机词不如有针对性的词来的有效，所以这里都是从全部三个分类里找到的词汇
1、创建词向量：中超/亚冠/国足/足协/英超/西甲/欧冠/意甲/德甲/篮球/NBA/CBA/高尔夫/乒乓/排球/网球/羽毛球/跑步/赛车/棋牌/台球/游泳/马术/拳击/田径/功夫/扑克/体育/球队/球员/训练/国家队/联赛/俱乐部/场地/翻盘/绝杀/热身/队友/冠军/亚军/季军/犯规/赛季/加时/反超/半场/争夺/战术/阵容/比赛/德比/恢复/进球/失球/奥斯卡/娱乐/影迷/电影/电视/音乐/戏剧/视频/演员/导演/明星/经纪人/歌手/连续剧/展映/粉丝/写真/演技/作秀/节目/艺人/超模/女星/模特/男星/性感/主创/院线/影业/拍摄/编剧/情节/影像/剧情/主演/上映/票房/开机/剧集/表演/收视/预告片/主持人/艾美奖/角色/剧院/乐迷/影迷/演出/专辑/乐坛/剧场/文艺/芭蕾/戏曲/舞蹈/军事/军队/军机/炸弹/军方/坦克/军舰/炸死/军演/战备/部队/军区/国防/士兵/舰船/潜艇/飞机/直升机/舰队/保卫/演习/武器/反击/打击/阅兵/对抗/防卫/海军/空军/陆军/武装/战略/空袭/冲突/装甲/步兵/作战/导弹/边防/侦察/战斗机/雷达/轰炸/防御/据点/火力/航空母舰/进攻/弹药/军营/包围/攻占/俘虏/參战/战友/战斗/入侵
2、搜狐上下载三类文章各10篇组成训练样本，计算出每篇文章的文档矩阵。标注每篇文章的类别标签
样本文件名称格式： 编号_类别标签.txt
文档矩阵：
000000000000000000100000000000000000001100010001001010000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000010000000000000100000000000000011110001010000000000000000011000000110000000000100000000000000000010000000000000000000000000000000000000000000000
000000000000000000000000000011000000000000000000000001001000001001000000001000000000000000000000010000000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000001001000000000000000001000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000010000000000000000000000000000010010000100000000000000010010000001000000000100000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000010000000010000010000010100000000111111111110000000100000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000110000000000000011010000001000010000000000001100001110000000000000000000000000000000000000000000000000000000000000000000000000000
000000000100000000000000000000000000000000000000000000001010000110000000000000000100000001101000000100000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000010000010000000000000001000000001100000100000000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000001010000110000000000000000000001011000010000110000000000000000000000000000000000000000000000000000000000000000000
000000010000000000000000000011100000001000010110001001000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000100000000000000000000000
000000000000000000000000000000000000000000000000000000001001000100000000000000000000000010000100000100000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000011110000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000110000000111111111111100000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001100000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000100000000000000000000100000000000000000000000000000000000
000000000000000000000000000000000000000000000001000000000001000000000000000000000000100000000000000000000000000100100000010010000000000000000100000000000100000000000010
000000000000000000000000000000100000000000000000000000000001000000000000000000000000000000000000000000000000000100010000010000000000000000000100000100000000000000000000
000000000000000000000000000000000000000000000000100000000000000000000000000000000000000000000000000000000000000000000000110010000000001001010000000010000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000000000100000000000000000000000000000010100000000000100000000010000000000000001000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001000010000000000100000000100010000000000001000000000
000000010000000000000000000111001100000000010000001001000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000100000000000100000000000000100000000110000010000000000000
110000000000000000000000000100001000100000010000000001000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001000000000
110000000000000000000000000111001100100100010001111011000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001000000000
000000000001000000000000000001000000101000100110001000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001000000000
000000000000010000000000000000000000001101000001001000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
000000000010000000000000000010000000000000010010001000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
000000000001000000000000000111100000101000110100001000100000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000001000000000
000000000000000000000000000100000000000110010100000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
类别标签向量：
122222222212333333333131111111
using System;
using System.Text;
using System.Windows.Forms;
using System.IO;
namespace NaiveBayes
{
public partial class Form1 : Form
{
private string[] vocabArray;
private double[] p0Num, p1Num, p2Num;
public Form1()
{
InitializeComponent();
label2.Text = "体育1、娱乐2、军事3\r\n每一个类型10个训练样本\r\n文章所有出自搜狐新闻\r\n词向量从各类文章中分词获得";
StreamReader sr = new StreamReader("vocabList.txt", Encoding.Default);
string line, all = "";
while ((line = sr.ReadLine()) != null)
{
all += line;
}
vocabArray = all.Split(new string[] { "/" }, StringSplitOptions.RemoveEmptyEntries);
}
private void Form1_Resize(object sender, EventArgs e)
{
this.Width = 800;
this.Height = 600;
}
private void button1_Click(object sender, EventArgs e)
{
//生成文档矩阵和分类标签向量
DirectoryInfo di = new DirectoryInfo("train");
FileInfo[] fi = di.GetFiles("*.txt");
string[] trainMatrix = new string[fi.Length];
p0Num = new double[vocabArray.Length];
p1Num = new double[vocabArray.Length];
p2Num = new double[vocabArray.Length];
double p0Denom = 2.0;
double p1Denom = 2.0;
double p2Denom = 2.0;
for (int i = 0; i < vocabArray.Length; i++)
{
p0Num[i] = p1Num[i] = p2Num[i] = 1.0;
}
string trainCategory = "";
int m = 0;
foreach (FileInfo i in fi)
{
StreamReader sr = new StreamReader(i.FullName, Encoding.Default);
string line, all = "";
while ((line = sr.ReadLine()) != null)
{
all += line;
}
string strVec = "";
foreach (string j in vocabArray)
{
if (all.Contains(j))
strVec += "1";
else
strVec += "0";
}
trainMatrix[m] = strVec;
m++;
trainCategory += i.Name.Substring(i.Name.LastIndexOf("_") + 1, 1);
}
StreamWriter sw = new StreamWriter(".\\trainV\\trainMatrix.txt", true);
foreach (string i in trainMatrix)
{
sw.WriteLine(i);
sw.Flush();
}
sw.Close();
sw = new StreamWriter(".\\trainV\\trainCategory.txt", true);
sw.WriteLine(trainCategory);
sw.Close();
for (int i = 0; i < trainMatrix.Length; i++)
{
if (trainCategory.Substring(i, 1) == "1")
{
double tmp = 0;
for (int j = 0; j < vocabArray.Length; j++)
{
p0Num[j] += double.Parse(trainMatrix[i].Substring(j, 1));
tmp += double.Parse(trainMatrix[i].Substring(j, 1));
}
p0Denom += tmp;
}
else if (trainCategory.Substring(i, 1) == "2")
{
double tmp = 0;
for (int j = 0; j < vocabArray.Length; j++)
{
p1Num[j] += double.Parse(trainMatrix[i].Substring(j, 1));
tmp += double.Parse(trainMatrix[i].Substring(j, 1));
}
p1Denom += tmp;
}
else if (trainCategory.Substring(i, 1) == "3")
{
double tmp = 0;
for (int j = 0; j < vocabArray.Length; j++)
{
p2Num[j] += double.Parse(trainMatrix[i].Substring(j, 1));
tmp += double.Parse(trainMatrix[i].Substring(j, 1));
}
p2Denom += tmp;
}
else
{
//Undo
}
}
for (int j = 0; j < vocabArray.Length; j++)
{
p0Num[j] = Math.Log(p0Num[j] / p0Denom);
p1Num[j] = Math.Log(p1Num[j] / p1Denom);
p2Num[j] = Math.Log(p2Num[j] / p2Denom);
}
label4.Text = "处理样本数据完毕";
}
private void button2_Click(object sender, EventArgs e)
{
if (textBox1.Text.Trim() != "")
{
string strVec = "";
foreach (string i in vocabArray)
{
if (textBox1.Text.Contains(i))
strVec += "1";
else
strVec += "0";
}
double p0 = 0;
double p1 = 0;
double p2 = 0;
for (int j = 0; j < vocabArray.Length; j++)
{
p0 += p0Num[j] * double.Parse(strVec.Substring(j, 1));
p1 += p1Num[j] * double.Parse(strVec.Substring(j, 1));
p2 += p2Num[j] * double.Parse(strVec.Substring(j, 1));
}
string catelog = "";
if (p0 > p1 && p0 > p2)
catelog = "体育";
else if (p1 > p0 && p1 > p2)
catelog = "娱乐";
else if (p2 > p0 && p2 > p1)
catelog = "军事";
else
catelog = "无法推断";
label3.Text = "体育：" + p0.ToString() + "\r\n娱乐：" + p1.ToString() + "\r\n军事：" + p2.ToString();
label1.Text = "所属类型是：" + catelog;
}
}
}
}
