In pattern recognition, K-mean algorithm is mainly used for clustering with known classification number, the realization is simple, the algorithm is clear, and belongs to one of the simpler dynamic clustering algorithms.
In the algorithm, the convergence of the clustering centers of the two-times algorithm is determined by iteration, and then the iteration is decided whether to continue the iterative (the algorithm is exited and the classification is completed).
A simple example of the following K-mean algorithm:
#Clustering Analysis of K-means algorithm fromNumPyImport* fromMathImportsqrt#requires a program to cluster the following data, taking K to 2x=array ([[[0,0],[1,0],[0,1],[1,1],[2,1], [1,2],[2,2],[3,2],[6,6],[7,6], [8,6],[6,7],[7,7],[8,7],[9,7], [7,8],[8,8],[9,8],[8,9],[9,9]])#finding the distance between the sample point and the cluster centerdefDis (b):returnsqrt ((a[0]-b[0]) **2+ (a[1]-b[1]) **2) #Classification of SamplesdefDistribution (X,Z_1,Z_2,N1,N2): forIinchRange (0,len (x)):if(Dis (x[i],z_1) <dis (x[i],z_2)): N1.append (i)Else: N2.append (i)#Creating a new cluster centerdefNew_dis_center (x,z_1,z_2,n1,n2): Z_1_1=[0,0] Z_2_1=[0,0] forIinchRange (0, (Len (N1))): Z_1_1=z_1_1+x[n1[i]]/Len (N1) forIinchRange (0, (Len (N2))): Z_2_1=z_2_1+x[n2[i]]/Len (n2)returnZ_1_1,z_2_1#iterate function input data and initial cluster centerdefIteration_fun (x,z_1,z_2):#Initialize the intermediate amount of the iterative cluster centerz_1_1=[0,0] Z_2_1=[0,0] N1=[] N2=[] Distribution (X,Z_1,Z_2,N1,N2) (z_1_1,z_2_1)=New_dis_center (X,Z_1,Z_2,N1,N2)#Iterative judgment if(z_1_1 = = z_1). All ()or(z_2_1==z_2). All ():Print("The clustering Center after iteration:") Print("Clustering Center One:") Print(z_1)Print("Clustering Center:") Print(z_2)Else: Z_1=z_1_1 z_2=z_2_1 Iteration_fun (x,z_1,z_2)#initializing cluster centers (including alternate iterative clustering centers) cluster Sample setz_1=x[0]z_2=x[1]iteration_fun (x,z_1,z_2)
Python--k-mean-algorithm the classification of sample points