Python根据成绩分析系统浅析

案例:该数据集的是一个关于每个学生成绩的数据集,接下来我们对该数据集进行分析，判断学生是否适合继续深造

数据集特征展示

1  gre 成绩 (290 to 340)
2  toefl 成绩(92 to 120)
3  学校等级 (1 to 5)
4  自身的意愿 (1 to 5)
5  推荐信的力度 (1 to 5)
6  cgpa成绩 (6.8 to 9.92)
7  是否有研习经验 (0 or 1)
8  读硕士的意向 (0.34 to 0.97)

1.导入包

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import os,sys

2.导入并查看数据集

df = pd.read_csv("d:\\machine-learning\\score\\admission_predict.csv",sep = ",")
print('there are ',len(df.columns),'columns')
for c in df.columns:
 sys.stdout.write(str(c)+', '

there are 9 columns
serial no., gre score, toefl score, university rating, sop, lor , cgpa, research, chance of admit , 
一共有9列特征

df.info()

<class 'pandas.core.frame.dataframe'>
rangeindex: 400 entries, 0 to 399
data columns (total 9 columns):
serial no.   400 non-null int64
gre score   400 non-null int64
toefl score   400 non-null int64
university rating 400 non-null int64
sop     400 non-null float64
lor     400 non-null float64
cgpa     400 non-null float64
research    400 non-null int64
chance of admit  400 non-null float64
dtypes: float64(4), int64(5)
memory usage: 28.2 kb

数据集信息：
1.数据有9个特征，分别是学号，gre分数，托福分数，学校等级，sop，lor，cgpa，是否参加研习，进修的几率
2.数据集中没有空值
3.一共有400条数据

# 整理列名称
df = df.rename(columns={'chance of admit ':'chance of admit'})
# 显示前5列数据
df.head()

3.查看每个特征的相关性

fig,ax = plt.subplots(figsize=(10,10))
sns.heatmap(df.corr(),ax=ax,annot=true,linewidths=0.05,fmt='.2f',cmap='magma')
plt.show()

结论：1.最有可能影响是否读硕士的特征是gre，cgpa，toefl成绩

2.影响相对较小的特征是lor，sop，和research

4.数据可视化，双变量分析

4.1 进行research的人数

print("not having research:",len(df[df.research == 0]))
print("having research:",len(df[df.research == 1]))
y = np.array([len(df[df.research == 0]),len(df[df.research == 1])])
x = np.arange(2)
plt.bar(x,y)
plt.title("research experience")
plt.xlabel("canditates")
plt.ylabel("frequency")
plt.xticks(x,('not having research','having research'))
plt.show()

　　结论：进行research的人数是219，本科没有research人数是181

　　4.2 学生的托福成绩

y = np.array([df['toefl score'].min(),df['toefl score'].mean(),df['toefl score'].max()])
x = np.arange(3)
plt.bar(x,y)
plt.title('toefl score')
plt.xlabel('level')
plt.ylabel('toefl score')
plt.xticks(x,('worst','average','best'))
plt.show()

结论：最低分92分，最高分满分，进修学生的英语成绩很不错

4.3 gre成绩

df['gre score'].plot(kind='hist',bins=200,figsize=(6,6))
plt.title('gre score')
plt.xlabel('gre score')
plt.ylabel('frequency')
plt.show()

结论：310和330的分值的学生居多

4.4 cgpa和学校等级的关系

plt.scatter(df['university rating'],df['cgpa'])
plt.title('cgpa scores for university ratings')
plt.xlabel('university rating')
plt.ylabel('cgpa')
plt.show()

结论：学校越好，学生的gpa可能就越高

4.5 gre成绩和cgpa的关系

plt.scatter(df['gre score'],df['cgpa'])
plt.title('cgpa for gre scores')
plt.xlabel('gre score')
plt.ylabel('cgpa')
plt.show()

结论：gpa基点越高，gre分数越高，2者的相关性很大

4.6 托福成绩和gre成绩的关系

df[df['cgpa']>=8.5].plot(kind='scatter',x='gre score',y='toefl score',color='red')
plt.xlabel('gre score')
plt.ylabel('toefl score')
plt.title('cgpa >= 8.5')
plt.grid(true)
plt.show()

结论：多数情况下gre和托福成正相关，但是gre分数高，托福一定高。

4.6 学校等级和是否读硕士的关系

s = df[df['chance of admit'] >= 0.75]['university rating'].value_counts().head(5)
plt.title('university ratings of candidates with an 75% acceptance chance')
s.plot(kind='bar',figsize=(20,10),cmap='pastel1')
plt.xlabel('university rating')
plt.ylabel('candidates')
plt.show()

结论：排名靠前的学校的学生，进修的可能性更大

4.7 sop和gpa的关系

plt.scatter(df['cgpa'],df['sop'])
plt.xlabel('cgpa')
plt.ylabel('sop')
plt.title('sop for cgpa')
plt.show()

结论： gpa很高的学生，选择读硕士的自我意愿更强烈

4.8 sop和gre的关系

plt.scatter(df['gre score'],df['sop'])
plt.xlabel('gre score')
plt.ylabel('sop')
plt.title('sop for gre score')
plt.show()

结论：读硕士意愿强的学生，gre分数较高

5.模型

5.1 准备数据集

# 读取数据集
df = pd.read_csv('d:\\machine-learning\\score\\admission_predict.csv',sep=',')

serialno = df['serial no.'].values

df.drop(['serial no.'],axis=1,inplace=true)
df = df.rename(columns={'chance of admit ':'chance of admit'})

# 分割数据集
y = df['chance of admit'].values
x = df.drop(['chance of admit'],axis=1)

from sklearn.model_selection import train_test_split
x_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.2,random_state=42)

# 归一化数据
from sklearn.preprocessing import minmaxscaler
scalex = minmaxscaler(feature_range=[0,1])
x_train[x_train.columns] = scalex.fit_transform(x_train[x_train.columns])
x_test[x_test.columns] = scalex.fit_transform(x_test[x_test.columns])

5.2 回归

5.2.1 线性回归

from sklearn.linear_model import linearregression

lr = linearregression()
lr.fit(x_train,y_train)
y_head_lr = lr.predict(x_test)

print('real value of y_test[1]: '+str(y_test[1]) + ' -> predict value: ' + str(lr.predict(x_test.iloc[[1],:])))
print('real value of y_test[2]: '+str(y_test[2]) + ' -> predict value: ' + str(lr.predict(x_test.iloc[[2],:])))

from sklearn.metrics import r2_score
print('r_square score: ',r2_score(y_test,y_head_lr))
y_head_lr_train = lr.predict(x_train)
print('r_square score(train data):',r2_score(y_train,y_head_lr_train))

5.2.2 随机森林回归

from sklearn.ensemble import randomforestregressor

rfr = randomforestregressor(n_estimators=100,random_state=42)
rfr.fit(x_train,y_train)
y_head_rfr = rfr.predict(x_test)

print('real value of y_test[1]: '+str(y_test[1]) + ' -> predict value: ' + str(rfr.predict(x_test.iloc[[1],:])))
print('real value of y_test[2]: '+str(y_test[2]) + ' -> predict value: ' + str(rfr.predict(x_test.iloc[[2],:])))

from sklearn.metrics import r2_score
print('r_square score: ',r2_score(y_test,y_head_rfr))
y_head_rfr_train = rfr.predict(x_train)
print('r_square score(train data):',r2_score(y_train,y_head_rfr_train))

5.2.3 决策树回归

from sklearn.tree import decisiontreeregressor

dt = decisiontreeregressor(random_state=42)
dt.fit(x_train,y_train)
y_head_dt = dt.predict(x_test)

print('real value of y_test[1]: '+str(y_test[1]) + ' -> predict value: ' + str(dt.predict(x_test.iloc[[1],:])))
print('real value of y_test[2]: '+str(y_test[2]) + ' -> predict value: ' + str(dt.predict(x_test.iloc[[2],:])))

from sklearn.metrics import r2_score
print('r_square score: ',r2_score(y_test,y_head_dt))
y_head_dt_train = dt.predict(x_train)
print('r_square score(train data):',r2_score(y_train,y_head_dt_train))

5.2.4 三种回归方法比较

y = np.array([r2_score(y_test,y_head_lr),r2_score(y_test,y_head_rfr),r2_score(y_test,y_head_dt)])
x = np.arange(3)
plt.bar(x,y)
plt.title('comparion of regression algorithms')
plt.xlabel('regression')
plt.ylabel('r2_score')
plt.xticks(x,("linearregression","randomforestreg.","decisiontreereg."))
plt.show()

结论 ： 回归算法中，线性回归的性能更优

5.2.5 三种回归方法与实际值的比较

​red = plt.scatter(np.arange(0,80,5),y_head_lr[0:80:5],color='red')
blue = plt.scatter(np.arange(0,80,5),y_head_rfr[0:80:5],color='blue')
green = plt.scatter(np.arange(0,80,5),y_head_dt[0:80:5],color='green')
black = plt.scatter(np.arange(0,80,5),y_test[0:80:5],color='black')
plt.title('comparison of regression algorithms')
plt.xlabel('index of candidate')
plt.ylabel('chance of admit')
plt.legend([red,blue,green,black],['lr','rfr','dt','real'])
plt.show()

结论：在数据集中有70%的候选人有可能读硕士，从上图来看还有些点没有很好的得到预测

5.3 分类算法

5.3.1 准备数据

df = pd.read_csv('d:\\machine-learning\\score\\admission_predict.csv',sep=',')

serialno = df['serial no.'].values
df.drop(['serial no.'],axis=1,inplace=true)

df = df.rename(columns={'chance of admit ':'chance of admit'})
y = df['chance of admit'].values
x = df.drop(['chance of admit'],axis=1)


from sklearn.model_selection import train_test_split
x_train,x_test,y_train,y_test = train_test_split(x,y,test_size=0.2,random_state=42)

from sklearn.preprocessing import minmaxscaler
scalex = minmaxscaler(feature_range=[0,1])
x_train[x_train.columns] = scalex.fit_transform(x_train[x_train.columns])
x_test[x_test.columns] = scalex.fit_transform(x_test[x_test.columns])

# 如果chance >0.8, chance of admit 就是1，否则就是0
y_train_01 = [1 if each > 0.8 else 0 for each in y_train]
y_test_01 = [1 if each > 0.8 else 0 for each in y_test]

y_train_01 = np.array(y_train_01)
y_test_01 = np.array(y_test_01)

5.3.2 逻辑回归

from sklearn.linear_model import logisticregression

lrc = logisticregression()
lrc.fit(x_train,y_train_01)
print('score: ',lrc.score(x_test,y_test_01))
print('real value of y_test_01[1]: '+str(y_test_01[1]) + ' -> predict value: ' + str(lrc.predict(x_test.iloc[[1],:])))
print('real value of y_test_01[2]: '+str(y_test_01[2]) + ' -> predict value: ' + str(lrc.predict(x_test.iloc[[2],:])))

from sklearn.metrics import confusion_matrix
cm_lrc = confusion_matrix(y_test_01,lrc.predict(x_test))

f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_lrc,annot=true,linewidths=0.5,linecolor='red',fmt='.0f',ax=ax)
plt.title('test for test dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

from sklearn.metrics import recall_score,precision_score,f1_score
print('precision_score is : ',precision_score(y_test_01,lrc.predict(x_test)))
print('recall_score is : ',recall_score(y_test_01,lrc.predict(x_test)))
print('f1_score is : ',f1_score(y_test_01,lrc.predict(x_test)))

# test for train dataset:

cm_lrc_train = confusion_matrix(y_train_01,lrc.predict(x_train))
f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_lrc_train,annot=true,linewidths=0.5,linecolor='blue',fmt='.0f',ax=ax)
plt.title('test for train dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

结论：1.通过混淆矩阵，逻辑回归算法在训练集样本上，有23个分错的样本，有72人想进一步读硕士

2.在测试集上有7个分错的样本　

5.3.3 支持向量机（svm）

from sklearn.svm import svc

svm = svc(random_state=1,kernel='rbf')
svm.fit(x_train,y_train_01)
print('score: ',svm.score(x_test,y_test_01))
print('real value of y_test_01[1]: '+str(y_test_01[1]) + ' -> predict value: ' + str(svm.predict(x_test.iloc[[1],:])))
print('real value of y_test_01[2]: '+str(y_test_01[2]) + ' -> predict value: ' + str(svm.predict(x_test.iloc[[2],:])))

from sklearn.metrics import confusion_matrix
cm_svm = confusion_matrix(y_test_01,svm.predict(x_test))

f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_svm,annot=true,linewidths=0.5,linecolor='red',fmt='.0f',ax=ax)
plt.title('test for test dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

from sklearn.metrics import recall_score,precision_score,f1_score
print('precision_score is : ',precision_score(y_test_01,svm.predict(x_test)))
print('recall_score is : ',recall_score(y_test_01,svm.predict(x_test)))
print('f1_score is : ',f1_score(y_test_01,svm.predict(x_test)))

# test for train dataset:

cm_svm_train = confusion_matrix(y_train_01,svm.predict(x_train))
f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_svm_train,annot=true,linewidths=0.5,linecolor='blue',fmt='.0f',ax=ax)
plt.title('test for train dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

结论：1.通过混淆矩阵，svm算法在训练集样本上，有22个分错的样本，有70人想进一步读硕士

2.在测试集上有8个分错的样本

5.3.4 朴素贝叶斯

from sklearn.naive_bayes import gaussiannb

nb = gaussiannb()
nb.fit(x_train,y_train_01)
print('score: ',nb.score(x_test,y_test_01))
print('real value of y_test_01[1]: '+str(y_test_01[1]) + ' -> predict value: ' + str(nb.predict(x_test.iloc[[1],:])))
print('real value of y_test_01[2]: '+str(y_test_01[2]) + ' -> predict value: ' + str(nb.predict(x_test.iloc[[2],:])))

from sklearn.metrics import confusion_matrix
cm_nb = confusion_matrix(y_test_01,nb.predict(x_test))

f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_nb,annot=true,linewidths=0.5,linecolor='red',fmt='.0f',ax=ax)
plt.title('test for test dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

from sklearn.metrics import recall_score,precision_score,f1_score
print('precision_score is : ',precision_score(y_test_01,nb.predict(x_test)))
print('recall_score is : ',recall_score(y_test_01,nb.predict(x_test)))
print('f1_score is : ',f1_score(y_test_01,nb.predict(x_test)))

# test for train dataset:

cm_nb_train = confusion_matrix(y_train_01,nb.predict(x_train))
f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_nb_train,annot=true,linewidths=0.5,linecolor='blue',fmt='.0f',ax=ax)
plt.title('test for train dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

结论：1.通过混淆矩阵，朴素贝叶斯算法在训练集样本上，有20个分错的样本，有78人想进一步读硕士

2.在测试集上有7个分错的样本

5.3.5 随机森林分类器

from sklearn.ensemble import randomforestclassifier

rfc = randomforestclassifier(n_estimators=100,random_state=1)
rfc.fit(x_train,y_train_01)
print('score: ',rfc.score(x_test,y_test_01))
print('real value of y_test_01[1]: '+str(y_test_01[1]) + ' -> predict value: ' + str(rfc.predict(x_test.iloc[[1],:])))
print('real value of y_test_01[2]: '+str(y_test_01[2]) + ' -> predict value: ' + str(rfc.predict(x_test.iloc[[2],:])))

from sklearn.metrics import confusion_matrix
cm_rfc = confusion_matrix(y_test_01,rfc.predict(x_test))

f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_rfc,annot=true,linewidths=0.5,linecolor='red',fmt='.0f',ax=ax)
plt.title('test for test dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

from sklearn.metrics import recall_score,precision_score,f1_score
print('precision_score is : ',precision_score(y_test_01,rfc.predict(x_test)))
print('recall_score is : ',recall_score(y_test_01,rfc.predict(x_test)))
print('f1_score is : ',f1_score(y_test_01,rfc.predict(x_test)))

# test for train dataset:

cm_rfc_train = confusion_matrix(y_train_01,rfc.predict(x_train))
f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_rfc_train,annot=true,linewidths=0.5,linecolor='blue',fmt='.0f',ax=ax)
plt.title('test for train dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

结论：1.通过混淆矩阵，随机森林算法在训练集样本上，有0个分错的样本，有88人想进一步读硕士

2.在测试集上有5个分错的样本

5.3.6 决策树分类器

from sklearn.tree import decisiontreeclassifier

dtc = decisiontreeclassifier(criterion='entropy',max_depth=3)
dtc.fit(x_train,y_train_01)
print('score: ',dtc.score(x_test,y_test_01))
print('real value of y_test_01[1]: '+str(y_test_01[1]) + ' -> predict value: ' + str(dtc.predict(x_test.iloc[[1],:])))
print('real value of y_test_01[2]: '+str(y_test_01[2]) + ' -> predict value: ' + str(dtc.predict(x_test.iloc[[2],:])))

from sklearn.metrics import confusion_matrix
cm_dtc = confusion_matrix(y_test_01,dtc.predict(x_test))

f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_dtc,annot=true,linewidths=0.5,linecolor='red',fmt='.0f',ax=ax)
plt.title('test for test dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

from sklearn.metrics import recall_score,precision_score,f1_score
print('precision_score is : ',precision_score(y_test_01,dtc.predict(x_test)))
print('recall_score is : ',recall_score(y_test_01,dtc.predict(x_test)))
print('f1_score is : ',f1_score(y_test_01,dtc.predict(x_test)))

# test for train dataset:

cm_dtc_train = confusion_matrix(y_train_01,dtc.predict(x_train))
f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_dtc_train,annot=true,linewidths=0.5,linecolor='blue',fmt='.0f',ax=ax)
plt.title('test for train dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

结论：1.通过混淆矩阵，决策树算法在训练集样本上，有20个分错的样本，有78人想进一步读硕士

2.在测试集上有7个分错的样本

5.3.7 k临近分类器

from sklearn.neighbors import kneighborsclassifier

scores = []
for each in range(1,50):
 knn_n = kneighborsclassifier(n_neighbors = each)
 knn_n.fit(x_train,y_train_01)
 scores.append(knn_n.score(x_test,y_test_01))
 
plt.plot(range(1,50),scores)
plt.xlabel('k')
plt.ylabel('accuracy')
plt.show()


knn = kneighborsclassifier(n_neighbors=7)
knn.fit(x_train,y_train_01)
print('score 7 : ',knn.score(x_test,y_test_01))
print('real value of y_test_01[1]: '+str(y_test_01[1]) + ' -> predict value: ' + str(knn.predict(x_test.iloc[[1],:])))
print('real value of y_test_01[2]: '+str(y_test_01[2]) + ' -> predict value: ' + str(knn.predict(x_test.iloc[[2],:])))

from sklearn.metrics import confusion_matrix
cm_knn = confusion_matrix(y_test_01,knn.predict(x_test))

f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_knn,annot=true,linewidths=0.5,linecolor='red',fmt='.0f',ax=ax)
plt.title('test for test dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

from sklearn.metrics import recall_score,precision_score,f1_score
print('precision_score is : ',precision_score(y_test_01,knn.predict(x_test)))
print('recall_score is : ',recall_score(y_test_01,knn.predict(x_test)))
print('f1_score is : ',f1_score(y_test_01,knn.predict(x_test)))

# test for train dataset:

cm_knn_train = confusion_matrix(y_train_01,knn.predict(x_train))
f,ax = plt.subplots(figsize=(5,5))
sns.heatmap(cm_knn_train,annot=true,linewidths=0.5,linecolor='blue',fmt='.0f',ax=ax)
plt.title('test for train dataset')
plt.xlabel('predicted y values')
plt.ylabel('real y value')
plt.show()

结论：1.通过混淆矩阵，k临近算法在训练集样本上，有22个分错的样本，有71人想进一步读硕士

2.在测试集上有7个分错的样本

5.3.8 分类器比较

y = np.array([lrc.score(x_test,y_test_01),svm.score(x_test,y_test_01),nb.score(x_test,y_test_01),
    dtc.score(x_test,y_test_01),rfc.score(x_test,y_test_01),knn.score(x_test,y_test_01)])
x = np.arange(6)
plt.bar(x,y)
plt.title('comparison of classification algorithms')
plt.xlabel('classification')
plt.ylabel('score')
plt.xticks(x,("logisticreg.","svm","gnb","dec.tree","ran.forest","knn"))
plt.show()

结论：随机森林和朴素贝叶斯二者的预测值都比较高

5.4 聚类算法

5.4.1 准备数据

df = pd.read_csv('d:\\machine-learning\\score\\admission_predict.csv',sep=',')
df = df.rename(columns={'chance of admit ':'chance of admit'})
serialno = df['serial no.']
df.drop(['serial no.'],axis=1,inplace=true)
df = (df - np.min(df)) / (np.max(df)-np.min(df))
y = df['chance of admit']
x = df.drop(['chance of admit'],axis=1)

5.4.2 降维

from sklearn.decomposition import pca

pca = pca(n_components=1,whiten=true)
pca.fit(x)
x_pca = pca.transform(x)
x_pca = x_pca.reshape(400)
dictionary = {'x':x_pca,'y':y}
data = pd.dataframe(dictionary)
print('pca data:',data.head())

print()

print('orin data:',df.head())

5.4.3 k均值聚类

from sklearn.cluster import kmeans

wcss = []
for k in range(1,15):
 kmeans = kmeans(n_clusters=k)
 kmeans.fit(x)
 wcss.append(kmeans.inertia_)
plt.plot(range(1,15),wcss)
plt.xlabel('kmeans')
plt.ylabel('wcss')
plt.show()

df["serial no."] = serialno
kmeans = kmeans(n_clusters=3)
clusters_knn = kmeans.fit_predict(x)
df['label_kmeans'] = clusters_knn


plt.scatter(df[df.label_kmeans == 0 ]["serial no."],df[df.label_kmeans == 0]['chance of admit'],color = "red")
plt.scatter(df[df.label_kmeans == 1 ]["serial no."],df[df.label_kmeans == 1]['chance of admit'],color = "blue")
plt.scatter(df[df.label_kmeans == 2 ]["serial no."],df[df.label_kmeans == 2]['chance of admit'],color = "green")
plt.title("k-means clustering")
plt.xlabel("candidates")
plt.ylabel("chance of admit")
plt.show()

plt.scatter(data.x[df.label_kmeans == 0 ],data[df.label_kmeans == 0].y,color = "red")
plt.scatter(data.x[df.label_kmeans == 1 ],data[df.label_kmeans == 1].y,color = "blue")
plt.scatter(data.x[df.label_kmeans == 2 ],data[df.label_kmeans == 2].y,color = "green")
plt.title("k-means clustering")
plt.xlabel("x")
plt.ylabel("chance of admit")
plt.show()

结论：数据集分成三个类别，一部分学生是决定继续读硕士，一部分放弃，还有一部分学生的比较犹豫，但是深造的可能性较大

5.4.4 层次聚类

from scipy.cluster.hierarchy import linkage,dendrogram

merg = linkage(x,method='ward')
dendrogram(merg,leaf_rotation=90)
plt.xlabel('data points')
plt.ylabel('euclidean distance')
plt.show()

from sklearn.cluster import agglomerativeclustering

hiyerartical_cluster = agglomerativeclustering(n_clusters=3,affinity='euclidean',linkage='ward')
clusters_hiyerartical = hiyerartical_cluster.fit_predict(x)
df['label_hiyerartical'] = clusters_hiyerartical

plt.scatter(df[df.label_hiyerartical == 0 ]["serial no."],df[df.label_hiyerartical == 0]['chance of admit'],color = "red")
plt.scatter(df[df.label_hiyerartical == 1 ]["serial no."],df[df.label_hiyerartical == 1]['chance of admit'],color = "blue")
plt.scatter(df[df.label_hiyerartical == 2 ]["serial no."],df[df.label_hiyerartical == 2]['chance of admit'],color = "green")
plt.title('hierarchical clustering')
plt.xlabel('candidates')
plt.ylabel('chance of admit')
plt.show()

plt.scatter(data[df.label_hiyerartical == 0].x,data.y[df.label_hiyerartical==0],color='red')
plt.scatter(data[df.label_hiyerartical == 1].x,data.y[df.label_hiyerartical==1],color='blue')
plt.scatter(data[df.label_hiyerartical == 2].x,data.y[df.label_hiyerartical==2],color='green')
plt.title('hierarchical clustering')
plt.xlabel('x')
plt.ylabel('chance of admit')
plt.show()

结论：从层次聚类的结果中，可以看出和k均值聚类的结果一致，只不过确定了聚类k的取值3

结论：通过本词入门数据集的训练，可以掌握

1.一些特征的展示的方法

2.如何调用sklearn 的api

3.如何取比较不同模型之间的好坏

代码+数据集：https://github.com/mounment/python-data-analyze/tree/master/kaggle/score