import pandas as pd
import numpy as np
import random
import math
import matplotlib.pyplot as plt
 
# 這兩行代碼解決 plt 中文顯示的問題
plt.rcParams['font.sans-serif'] = ['SimHei']
plt.rcParams['axes.unicode_minus'] = False
 
df = pd.read_excel('13信科學(xué)生成績(jī).xlsx')
data = np.array(df)
 
df.head(10)

# 輸入數(shù)據(jù)
x = data.T[0]
y = data.T[1]
 
plt.scatter(x, y, s=50, c='r')  # 畫散點(diǎn)圖
plt.xlabel('平時(shí)')  # 橫坐標(biāo)軸標(biāo)題
plt.ylabel('期末')  # 縱坐標(biāo)軸標(biāo)題
plt.show()

k=3
i = 1
 
min1 = data.min(axis = 0)
max1 = data.max(axis = 0)
 
#在數(shù)據(jù)最大最小值中隨機(jī)生成k個(gè)初始聚類中心，保存為t
centre = np.empty((k,2))
for i in range(k):
    centre[i][0] = random.randint(min1[0],max1[0])#平時(shí)成績(jī)
    centre[i][1] = random.randint(min1[1],max1[1])#期末成績(jī)
 
while i<500:
    
    #計(jì)算歐氏距離
    def euclidean_distance(List,t):
        return math.sqrt(((List[0] - t[0])**2 + (List[1] - t[1])**2))
 
    #每個(gè)點(diǎn)到每個(gè)中心點(diǎn)的距離矩陣
    dis = np.empty((len(data),k))
    for i in range(len(data)):
        for j in range(k):
            dis[i][j] = euclidean_distance(data[i],centre[j])
    
    #初始化分類矩陣
    classify = []
    for i in range(k):
        classify.append([])
    
    #比較距離并分類
    for i in range(len(data)):
        List = dis[i].tolist()
        index = List.index(dis[i].min())
        classify[index].append(i)
    
    #構(gòu)造新的中心點(diǎn)
    new_centre = np.empty((k,2))
    for i in range(len(classify)):
 
        new_centre[i][0] = np.sum(data[classify[i]][0])/len(classify[i])
        new_centre[i][1] = np.sum(data[classify[i]][1])/len(classify[i])
     
    #比較新的中心點(diǎn)和舊的中心點(diǎn)是否一樣
    if (new_centre == centre).all():
        break
    else:
        centre = new_centre
        i = i + 1
 
# print('迭代次數(shù)為：',i)
print('聚類中心為：',new_centre)
print('分類情況為：',classify)

mark = ['or', 'ob', 'og', 'ok','sb', 'db', '<b', 'pb'] #紅、藍(lán)、綠、黑四種顏色的圓點(diǎn)
#mark=['sb', 'db', '<b', 'pb']
plt.figure(3)#創(chuàng)建圖表1  
for i in range(0,k):
 
    x=[]
    y=[]
    for j in range(len(classify[i])):
        x.append(data[classify[i][j]][0])
        y.append(data[classify[i][j]][1])
 
    plt.xlim(xmax=105,xmin=45)
    plt.ylim(ymax=85,ymin=-5)
    plt.plot(x,y,mark[i])
    #plt.show()

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from sklearn.cluster import KMeans  
from sklearn import metrics
 
df = pd.read_excel('13信科學(xué)生成績(jī).xlsx')
data = np.array(df)
y_pred=KMeans(n_clusters=3,random_state=9).fit_predict(data)
plt.scatter(data[:,0],data[:,1],c=y_pred)
plt.show()
print(metrics.calinski_harabasz_score(data,y_pred))