吴恩达老师机器学习作业-ex7

导入库，读取数据，查看数据类型等进行分析，可视化数据

import matplotlib.pyplot as plt
import numpy as np
import scipy.io as sio#读取数据
path = "./ex7data2.mat"
data = sio.loadmat(path)
# print(type(data))
# print(data.keys())
X = data.get("X")
# print(X.shape)
# print(X)
# print(type(X))#可视化数据
plt.scatter(X[:,0],X[:,1])
plt.show()

簇分配

在该函数中，设置一个minis表示样本与各个聚类中心点的距离，所以最开始初始化为一个比较大的数值。第一个for循环为迭代每一个样本，第二个for循环迭代每一个聚类中心点，已便找到最近的聚类中心点。

def means_classification(X,centros):m = len(X)n = len(centros)idx = np.zeros(m)for i in range(m):minis = 100000for j in range(n):s = np.sum(np.power((X[i,:]-centros[j,:]),2))if s < minis:minis = sidx[i] = jreturn idx

计算聚类中心

使用平均值计算聚类中心

def means_center(X,K,idx):centors = []for i in range(K):indicates = np.where(idx == i)centors_i = np.mean(X[indicates],axis=0)centors.append(centors_i)return centors

初始化聚类中心

随机生成三个整数（样本范围内），然后当作下标索引，找到三个样本的点作为初始化的聚类中心。

def init_centros(X,K):m = len(X)random_int = []for i in range(K):random_int_i = np.random.randint(0, m)random_int.append(random_int_i)random_centros = []for k in random_int:random_centros.append(X[k])return np.array(random_centros)

手动定义聚类数量为3

K = 3

多次迭代，运行Kmeans算法

def run_Kmeans(X,K,times):for i in range(times):centros = init_centros(X, K)idx = means_classification(X, centros)centros = means_center(X, K, idx)return idx,centros

绘制出聚类算法后的散点图

def plot_kmeans(X,idx):cluster1 = X[np.where(idx == 0)[0], :]cluster2 = X[np.where(idx == 1)[0], :]cluster3 = X[np.where(idx == 2)[0], :]fig,ax = plt.subplots()ax.scatter(cluster1[:, 0], cluster1[:, 1], c="r", label="cluster1")ax.scatter(cluster2[:, 0], cluster2[:, 1], c="g", label="cluster2")ax.scatter(cluster3[:, 0], cluster3[:, 1], c="b", label="cluster3")plt.show()idx,centros = run_Kmeans(X,K,100)
plot_kmeans(X,idx)