import numpy as np  import matplotlib.pyplot as plt   w=250 train=np.random.randint(-300,300,(w,2)) train=train.astype(float) train_lable=np.zeros(w) for i in range(2):     train[:,i]=(train[:,i]-train[:,i].mean())/train[:,i].std() for i in range(250):     if train[i,0]>train[i,1]:         plt.scatter(train[i,0], train[i,1],c='b')         train_lable[i]=1     else:         plt.scatter(train[i,0], train[i,1],c='c')         train_lable[i]=-1 plt.axis('equal') plt.show() def knn(index,k):     dis=np.zeros(200)     disl=[]     for i in range(200):         dis[i]=(train[index,0]-train[i,0])*(train[index,0]-train[i,0])+(train[index,1]-train[i,1])*(train[index,1]-train[i,1])         disl.append(dis[i])     target=[]     for i in range(k):         target.append(disl.index(min(disl)))         disl[disl.index(min(disl))]=float('inf')     l1=0     l2=0     for i in range(k):         if train_lable[target[i]]==1:             l1=l1+1         else:             l2=l2+1     if l1>l2 :         return 1     else:         return -1
sum1=0 for i in range(50):     lable=knn(i+200,3)     if lable==train_lable[i+200]:         sum1=sum1+1 print(sum1)