数据挖掘的分类算法
数据分析师对数据进行数据挖掘时，会运用很多算法，其中一种是分类算法，下面就对数据分析师运用分类算法进行一下详解，如下：
| @@ -0,0 +1,15 @@ | ||
| +packageDataMining_HITS; | ||
| + | ||
| +/** | ||
| +* HITSÁ´½Ó·ÖÎöËã·¨ | ||
| +* @author lyq | ||
| +* | ||
| +*/ | ||
| +publicclassClient{ | ||
| +publicstaticvoidmain(String[]args){ | ||
| +StringfilePath="C:\\Users\\lyq\\Desktop\\icon\\input.txt"; | ||
| + | ||
| +HITSTooltool=newHITSTool(filePath); | ||
| + tool.printResultPage(); | ||
| + } | ||
| +} |
| @@ -0,0 +1,150 @@ | ||
| +packageDataMining_HITS; | ||
| + | ||
| +importjava.io.BufferedReader; | ||
| +importjava.io.File; | ||
| +importjava.io.FileReader; | ||
| +importjava.io.IOException; | ||
| +importjava.util.ArrayList; | ||
| + | ||
| +/** | ||
| +* HITS链接分析算法工具类 | ||
| +* @author lyq | ||
| +* | ||
| +*/ | ||
| +publicclassHITSTool{ | ||
| +//输入数据文件地址 | ||
| +privateStringfilePath; | ||
| +//网页个数 | ||
| +privateintpageNum; | ||
| +//网页Authority权威值 | ||
| +privatedouble[] authority; | ||
| +//网页hub中心值 | ||
| +privatedouble[] hub; | ||
| +//链接矩阵关系 | ||
| +privateint[][] linkMatrix; | ||
| +//网页种类 | ||
| +privateArrayList<String>pageClass; | ||
| + | ||
| +publicHITSTool(StringfilePath){ | ||
| +this.filePath=filePath; | ||
| + readDataFile(); | ||
| + } | ||
| + | ||
| +/** | ||
| +* 从文件中读取数据 | ||
| +*/ | ||
| +privatevoidreadDataFile() { | ||
| +Filefile=newFile(filePath); | ||
| +ArrayList<String[]>dataArray=newArrayList<String[]>(); | ||
| + | ||
| +try{ | ||
| +BufferedReaderin=newBufferedReader(newFileReader(file)); | ||
| +Stringstr; | ||
| +String[] tempArray; | ||
| +while((str=in.readLine())!=null) { | ||
| + tempArray=str.split(""); | ||
| + dataArray.add(tempArray); | ||
| + } | ||
| + in.close(); | ||
| + }catch(IOExceptione) { | ||
| + e.getStackTrace(); | ||
| + } | ||
| + | ||
| + pageClass=newArrayList<>(); | ||
| +// 统计网页类型种数 | ||
| +for(String[] array:dataArray) { | ||
| +for(Strings:array) { | ||
| +if(!pageClass.contains(s)) { | ||
| + pageClass.add(s); | ||
| + } | ||
| + } | ||
| + } | ||
| + | ||
| +inti=0; | ||
| +intj=0; | ||
| + pageNum=pageClass.size(); | ||
| + linkMatrix=newint[pageNum][pageNum]; | ||
| + authority=newdouble[pageNum]; | ||
| + hub=newdouble[pageNum]; | ||
| +for(intk=0; k<pageNum; k++){ | ||
| +//初始时默认权威值和中心值都为1 | ||
| + authority[k]=1; | ||
| + hub[k]=1; | ||
| + } | ||
| + | ||
| +for(String[] array:dataArray) { | ||
| + | ||
| + i=Integer.parseInt(array[0]); | ||
| + j=Integer.parseInt(array[1]); | ||
| + | ||
| +// 设置linkMatrix[i][j]为1代表i网页包含指向j网页的链接 | ||
| + linkMatrix[i-1][j-1]=1; | ||
| + } | ||
| + } | ||
| + | ||
| +/** | ||
| +* 输出结果页面，也就是authority权威值最高的页面 | ||
| +*/ | ||
| +publicvoidprintResultPage(){ | ||
| +//最大Hub和Authority值，用于后面的归一化计算 | ||
| +doublemaxHub=0; | ||
| +doublemaxAuthority=0; | ||
| +intmaxAuthorityIndex=0; | ||
| +//误差值，用于收敛判断 | ||
| +doubleerror=Integer.MAX_VALUE; | ||
| +double[] newHub=newdouble[pageNum]; | ||
| +double[] newAuthority=newdouble[pageNum]; | ||
| + | ||
| + | ||
| +while(error>0.01*pageNum){ | ||
| +for(intk=0; k<pageNum; k++){ | ||
| + newHub[k]=0; | ||
| + newAuthority[k]=0; | ||
| + } | ||
| + | ||
| +//hub和authority值的更新计算 | ||
| +for(inti=0; i<pageNum; i++){ | ||
| +for(intj=0; j<pageNum; j++){ | ||
| +if(linkMatrix[i][j]==1){ | ||
| + newHub[i]+=authority[j]; | ||
| + newAuthority[j]+=hub[i]; | ||
| + } | ||
| + } | ||
| + } | ||
| + | ||
| + maxHub=0; | ||
| + maxAuthority=0; | ||
| +for(intk=0; k<pageNum; k++){ | ||
| +if(newHub[k]>maxHub){ | ||
| + maxHub=newHub[k]; | ||
| + } | ||
| + | ||
| +if(newAuthority[k]>maxAuthority){ | ||
| + maxAuthority=newAuthority[k]; | ||
| + maxAuthorityIndex=k; | ||
| + } | ||
| + } | ||
| + | ||
| + error=0; | ||
| +//归一化处理 | ||
| +for(intk=0; k<pageNum; k++){ | ||
| + newHub[k]/=maxHub; | ||
| + newAuthority[k]/=maxAuthority; | ||
| + | ||
| + error+=Math.abs(newHub[k]-hub[k]); | ||
| +System.out.println(newAuthority[k]+":"+newHub[k]); | ||
| + | ||
| + hub[k]=newHub[k]; | ||
| + authority[k]=newAuthority[k]; | ||
| + } | ||
| +System.out.println("---------"); | ||
| + } | ||
| + | ||
| +System.out.println("****最终收敛的网页的权威值和中心值****"); | ||
| +for(intk=0; k<pageNum; k++){ | ||
| +System.out.println("网页"+pageClass.get(k)+":"+authority[k]+":"+hub[k]); | ||
| + } | ||
| +System.out.println("权威值最高的网页为：网页"+pageClass.get(maxAuthorityIndex)); | ||
| + } | ||
| + | ||
| +} |
| @@ -0,0 +1,4 @@ | ||
| +1 2 | ||
| +1 3 | ||
| +2 3 | ||
| +3 1 |
数据分析咨询请扫描二维码
若不方便扫码，搜微信号：CDAshujufenxi
京公网安备 11010802034615号
经营许可证编号：京B2-20210330
