diff --git "a/\344\275\234\344\270\2323-\344\275\225\350\231\216\344\274\237" "b/\344\275\234\344\270\2323-\344\275\225\350\231\216\344\274\237"
new file mode 100644
index 0000000000000000000000000000000000000000..211b823a4a99784eb48e7dd53ae86ccdf0396131
--- /dev/null
+++ "b/\344\275\234\344\270\2323-\344\275\225\350\231\216\344\274\237"
@@ -0,0 +1,102 @@
+import pandas as pd
+import random
+import numpy as np
+import matplotlib.pyplot as plt
+import math
+
+path = "dataset_circles.csv"
+dataset = pd.read_csv(path)
+data_array = np.array(dataset)# 首先将pandas读取的数据转化为array
+data_list =data_array.tolist()# 然后转化为list形式
+data_list=np.delete(data_list,2,axis=1)
+a=b=n=0
+for i in range(len(data_list)):
+    a=data_list[i][0]+a
+    b=data_list[i][1]+b
+    n=n+1
+avx=a/n
+avy=b/n
+av=[avx,avy]
+def change(X):
+    n=np.shape(X)[0]
+    A=np.zeros((n,2))
+    for i in range(n):
+        A[i][0]=math.sqrt(X[i][0]*X[i][0]+X[i][1]*X[i][1])
+        A[i][1]=math.atan(X[i][1]/X[i][0])
+    return A
+def change1(X):
+    n=np.shape(X)[0]
+    A=np.zeros((n,2))
+    for i in range(n):
+        A[i][0]=math.sqrt(pow(X[i][0]-avx,2)+pow(X[i][1]-avy,2))
+        A[i][1]=math.atan(((X[i][1]-avy)/(X[i][0]-avx)))
+def distance(x1, x2):
+    return  abs(x2[0]-x1[0])+abs(x2[1]-x1[1])
+def rand_cluster_cents(X, k):  
+    m=np.shape(X)[0]    # 样本数 
+    dataIndex=list(range(m))    # 生成随机下标列表
+    random.shuffle(dataIndex)#打乱列表
+    centroidsIndex = dataIndex[:k]     
+    return X[centroidsIndex, :]# 返回随机的聚类中心
+def kmeans(X, k): 
+    m = np.shape(X)[0]# 样本总数
+    clusterAssment = np.zeros((m, 2)) # 分配样本到最近的簇：存[簇序号,距离的平方] (m行 x 2 列)
+    centroids = rand_cluster_cents(X, k)# step1: 通过随机产生的样本点初始化聚类中心
+    print('最初的中心=', centroids)
+    iterN = 0   
+    while True:   
+        clusterChanged = False 
+        for i in range(m):# step2:分配到最近的聚类中心对应的簇中
+            minDist = np.inf# minDist值表示+∞
+            minIndex = -1
+            for j in range(k):# 计算第i个样本到第j个中心点的距离
+                distJI = distance(centroids[j], X[i])
+                if distJI < minDist:
+                    minDist = distJI
+                    minIndex = j                      
+            if clusterAssment[i, 0] != minIndex:# 样本上次分配结果跟本次不一样，标志位clusterChanged置True
+                clusterChanged = True
+            clusterAssment[i, :] = minIndex, minDist ** 2  # 分配样本到最近的簇          
+        iterN += 1
+        sse = sum(clusterAssment[:, 1])
+        print('the SSE of %d' % iterN + 'th iteration is %f' % sse)
+        for cent in range(k):  # step3:更新聚类中心 # 样本分配结束后，重新计算聚类中心
+            ptsInClust = X[clusterAssment[:, 0] == cent, :]
+            centroids[cent, :] = np.mean(ptsInClust, axis=0)
+        if not clusterChanged: # 如果聚类重心没有发生改变，则退出迭代
+            break           
+    return centroids, clusterAssment
+"""
+# 进行k-means聚类
+"""
+k = 2  # 定义聚类数
+def datashow(dataSet, k, centroids, clusterAssment):  # 二维空间显示聚类结果
+    num, dim = np.shape(dataSet)  # 样本数num ,维数dim
+    if dim != 2:
+        print('sorry,the dimension of your dataset is not 2!')
+        return 1
+    marksamples = ['or', 'ob', 'og', 'ok', '^r', '^b', '<g']  # 样本图形标记
+    if k > len(marksamples):
+        print('sorry,your k is too large,please add length of the marksample!')
+        return 1
+    for i in range(num):# 绘所有样本
+        markindex = int(clusterAssment[i, 0])  # 矩阵形式转为int值, 簇序号
+        plt.plot(dataSet[i, 0], dataSet[i, 1], marksamples[markindex], markersize=6) # 特征维对应坐标轴x,y；样本图形标记及大小
+    markcentroids = ['o', '^']   # 绘中心点,聚类中心图形标记
+    label = ['0', '1']
+    c = ['yellow', 'red']
+    for i in range(k):
+        plt.plot(centroids[i, 0], centroids[i, 1], markcentroids[i], markersize=15, label=label[i], c=c[i])
+        plt.legend(loc='upper left')  #图例
+    plt.xlabel('feature 1')
+    plt.ylabel('feature 2')
+    plt.title(' result')  # 标题
+    plt.show()
+C=change(data_list)
+D=change(data_list)
+mycentroids, clusterAssment = kmeans(C, k)
+datashow(C, k, mycentroids, clusterAssment)
+mycentroids, clusterAssment = kmeans(data_list, k)
+datashow(data_list, k, mycentroids, clusterAssment)
+mycentroids, clusterAssment = kmeans(D, k)
+datashow(D, k, mycentroids, clusterAssment)