dbscan.zip_DBSCAN_dbscanpython实现_python DBSCAN_python实现dbscan_ra

# from sklearn import datasets import numpy as np #科学计算包 import random import matplotlib.pyplot as plt #2D绘图库 import time begin=time.time() def findNeighbor(j,X,eps): #eps是扫描半径 N=[] #令N为p的eps邻域中的对象的集合 for p in range(X.shape[0]): #找到所有领域内对象 temp=np.sqrt(np.sum(np.square(X[j]-X[p]))) #欧氏距离 if(temp<=eps): N.append(p) #将p加入N列表内 return N def dbscan(X,eps,min_Pts): k=-1 NeighborPts=[] #array,某点领域内的对象 Ner_NeighborPts=[] fil=[] #初始时已访问对象列表为空 gama=[x for x in range(len(X))] #初始时将所有点标记为未访问 cluster=[-1 for y in range(len(X))] while len(gama)>0: j=random.choice(gama) gama.remove(j) #未访问列表中移除 fil.append(j) #添加入访问列表 NeighborPts=findNeighbor(j,X,eps) if len(NeighborPts) < min_Pts: #比较N列表的长度和min_Pts cluster[j]=-1 #若关于j的列表的长度过小，将j标记为噪声点 else: k=k+1 cluster[j]=k #j不是噪声点 for i in NeighborPts: #访问N列表中的点 if i not in fil: #如果点未被添加到访问列表 gama.remove(i) #将点从未访问列表移除 fil.append(i) #将点添加到访问列表中 Ner_NeighborPts=findNeighbor(i,X,eps) if len(Ner_NeighborPts) >= min_Pts: #判断点是否也是核心对象 for a in Ner_NeighborPts: #访问核心对象i的密度可达点 if a not in NeighborPts: NeighborPts.append(a) #如果a不在N列表内，则进行添加 if (cluster[i]==-1): #-------------- cluster[i]=k print(k) return cluster #-------------------------------------------- # X1, y1=datasets.make_circles(n_samples=5000, factor=.6, # noise=.05) # X2, y2 = datasets.make_blobs(n_samples=1000, n_features=2, centers=[[1.2,1.2]], cluster_std=[[.1]], # random_state=9) # X = np.concatenate((X1, X2)) dataset="Iris1.txt" X=np.loadtxt(dataset, delimiter='\t') eps=1 min_Pts=1 C=dbscan(X,eps,min_Pts) plt.figure(num='dbscan',figsize=(12, 9), dpi=80, facecolor='pink',edgecolor='red', frameon=True) plt.scatter(X[:,0],X[:,1],marker='*',c=C) plt.show() end=time.time() print ("my dbscan用时:",end-begin)