package orisun;import java.io.File; import java.util.ArrayList; import java.util.Vector; import java.util.Iterator;public class DBScan {double Eps=3; //區(qū)域半徑int MinPts=4; //密度//由于自己到自己的距離是0,所以自己也是自己的neighborpublic Vector<DataObject> getNeighbors(DataObject p,ArrayList<DataObject> objects){Vector<DataObject> neighbors=new Vector<DataObject>();Iterator<DataObject> iter=objects.iterator();while(iter.hasNext()){DataObject q=iter.next();double[] arr1=p.getVector();double[] arr2=q.getVector();int len=arr1.length;if(Global.calEditDist(arr1,arr2,len)<=Eps){ //使用編輯距離 // if(Global.calEuraDist(arr1, arr2, len)<=Eps){ //使用歐氏距離 // if(Global.calCityBlockDist(arr1, arr2, len)<=Eps){ //使用街區(qū)距離 // if(Global.calSinDist(arr1, arr2, len)<=Eps){ //使用向量夾角的正弦 neighbors.add(q);}}return neighbors;}public int dbscan(ArrayList<DataObject> objects){int clusterID=0;boolean AllVisited=false;while(!AllVisited){Iterator<DataObject> iter=objects.iterator();while(iter.hasNext()){DataObject p=iter.next();if(p.isVisited())continue;AllVisited=false;p.setVisited(true); //設(shè)為visited后就已經(jīng)確定了它是核心點(diǎn)還是邊界點(diǎn)Vector<DataObject> neighbors=getNeighbors(p,objects);if(neighbors.size()<MinPts){if(p.getCid()<=0)p.setCid(-1); //cid初始為0,表示未分類；分類后設(shè)置為一個(gè)正數(shù)；設(shè)置為-1表示噪聲。}else{if(p.getCid()<=0){clusterID++;expandCluster(p,neighbors,clusterID,objects);}else{int iid=p.getCid();expandCluster(p,neighbors,iid,objects);}}AllVisited=true;}}return clusterID;}private void expandCluster(DataObject p, Vector<DataObject> neighbors,int clusterID,ArrayList<DataObject> objects) {p.setCid(clusterID);Iterator<DataObject> iter=neighbors.iterator();while(iter.hasNext()){DataObject q=iter.next();if(!q.isVisited()){q.setVisited(true);Vector<DataObject> qneighbors=getNeighbors(q,objects);if(qneighbors.size()>=MinPts){Iterator<DataObject> it=qneighbors.iterator();while(it.hasNext()){DataObject no=it.next();if(no.getCid()<=0)no.setCid(clusterID);}}}if(q.getCid()<=0){ //q不是任何簇的成員 q.setCid(clusterID);}}}public static void main(String[] args){DataSource datasource=new DataSource();//Eps=3,MinPts=4datasource.readMatrix(new File("/home/orisun/test/dot.mat"));datasource.readRLabel(new File("/home/orisun/test/dot.rlabel"));//Eps=2.5,MinPts=4 // datasource.readMatrix(new File("/home/orisun/text.normalized.mat")); // datasource.readRLabel(new File("/home/orisun/text.rlabel"));DBScan ds=new DBScan();int clunum=ds.dbscan(datasource.objects);datasource.printResult(datasource.objects,clunum);} }

?

轉(zhuǎn)載于:https://www.cnblogs.com/wukong0214/p/3440482.html

總結(jié)

以上是生活随笔為你收集整理的聚类算法之DBScan(Java实现)[转]的全部?jī)?nèi)容，希望文章能夠幫你解決所遇到的問(wèn)題。

如果覺(jué)得生活随笔網(wǎng)站內(nèi)容還不錯(cuò)，歡迎將生活随笔推薦給好友。