python数据分析与应用笔记

使用sklearn构建模型

1.使用sklearn转换器处理数据

import numpy as np
from sklearn.datasets import load_breast_cancer
from sklearn.preprocessing import MinMaxScaler  #该函数时对数据做标准化处理
from sklearn.decomposition import PCA  #该函数时对数据进行降维处理
from sklearn.model_selection import train_test_split  #该函数是对数据做训练集和测试集的划分
cancer = load_breast_cancer()   #将数据集赋值给cancer变量
cancer_data = cancer['data']   #提取数据集中的数据
cancer_target = cancer['target']   #提取数据集中的标签
cancer_names = cancer['feature_names']  #查看特征数目
cancer_desc = cancer['DESCR']

#划分训练集和测试集，其中20%的作为测试集
cancer_train_data,cancer_test_data,cancer_train_target,cancer_test_target = train_test_split(cancer_data,cancer_target,test_size = 0.2,random_state = 42)
scaler = MinMaxScaler().fit(cancer_train_data)  #生成规则
# 将规则应用于训练集和测试集
cancer_trainScaler = scaler.transform(cancer_train_data)
cancer_testScaler = scaler.transform(cancer_test_data)
#构建pca降维模型
pca_model = PCA(n_components = 10).fit(cancer_trainScaler)
#将降维模型应用于标准化之后的训练数据和测试数据
cancer_trainPca = pca_model.transform(cancer_trainScaler)
cancer_testPca = pca_model.transform(cancer_testScaler)

print('降维前训练数据的形状：',cancer_trainScaler.shape)
print('降维后训练数据的形状：',cancer_trainPca.shape)
print('降维前测试数据的形状：',cancer_testScaler.shape)
print('降维后测试数据的形状：',cancer_testPca.shape)

降维前训练数据的形状： (455, 30)
降维后训练数据的形状： (455, 10)
降维前测试数据的形状： (114, 30)
降维后测试数据的形状： (114, 10)

• 任务：使用sklearn实现数据处理和降维操作

from sklearn.datasets import load_boston
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.decomposition import PCA  
boston = load_boston()
boston_data = boston['data']
boston_target = boston['target']
boston_names = boston['feature_names']
boston_train_data,boston_test_data,boston_train_target,boston_test_target = train_test_split(boston_data,boston_target,test_size = 0.2,random_state = 42)
stdScale = StandardScaler().fit(boston_train_data)
boston_trainScaler = stdScale.transform(boston_train_data)
boston_testScaler =  stdScale.transform(boston_test_data)

pca_model = PCA(n_components = 5).fit(boston_trainScaler)
boston_trainPca = pca_model.transform(boston_trainScaler)
boston_testPca = pca_model.transform(boston_testScaler)

2.构建并评价聚类模型

常用的聚类算法如表所示：

sklearn常用的聚类算法模块cluster提供的聚类算法及其适用范围如图：

import pandas as pd
from sklearn.manifold import TSNE  #TSNE函数可实现多维数据的可视化展现
import matplotlib.pyplot as plt
from sklearn.datasets import load_iris
from sklearn.preprocessing import MinMaxScaler
from sklearn.cluster import KMeans 
iris = load_iris()
iris_data = iris['data']
iris_target = iris['target']
iris_names = iris['feature_names']
scale = MinMaxScaler().fit(iris_data)  #构建规则
iris_dataScale = scale.transform(iris_data)  #将规则应用于数据
kmeans = KMeans(n_clusters = 3,random_state = 123).fit(iris_dataScale) #构建并训练聚类模型
result = kmeans.predict([[1.5,1.5,1.5,1.5]])  #用模型进行预测

tsne = TSNE(n_components = 2,init = 'random',random_state=177).fit(iris_data)  #使用TSNE对数据进行降维，降成两维
df = pd.DataFrame(tsne.embedding_)  #将原始数据转化为DataFrame
df['labels']=kmeans.labels_  #将聚类结果存储进df数据集

df1 = df[df['labels']==0]
df2 = df[df['labels']==1]
df3 = df[df['labels']==2]

fig = plt.figure(figsize=(9,6))
plt.plot(df1[0],df1[1],'bo',df2[0],df2[1],'r*',df3[0],df3[1],'gD')
#plt.axis([-60,60,-80,80])
plt.savefig('聚类结果.png')
plt.show()
# print(df)
# print(df1)
# print(kmeans.labels_)
print(iris_names)

['sepal length (cm)', 'sepal width (cm)', 'petal length (cm)', 'petal width (cm)']

评价聚类模型

• 标准是：组内相似性越大，组间差别越大，其聚类效果越好
  sklearn 的metrics模块提供的聚类模型评价指标有：

使用FMI评级法去评价K-Means聚类模型

from sklearn.metrics import fowlkes_mallows_score
for i in range(2,7):
    kmeans = KMeans(n_clusters = i,random_state = 123).fit(iris_data)   
    score = fowlkes_mallows_score(iris_target,kmeans.labels_)
    print('iris数据聚%d类FMI评价分值为：%f'%(i,score))

iris数据聚2类FMI评价分值为：0.750473
iris数据聚3类FMI评价分值为：0.820808
iris数据聚4类FMI评价分值为：0.753970
iris数据聚5类FMI评价分值为：0.725483
iris数据聚6类FMI评价分值为：0.600691

使用轮廓系数评价法

from sklearn.metrics import silhouette_score
import matplotlib.pyplot as plt
silhouettteScore = []
for i in range(2,15):
    kmeans = KMeans(n_clusters = i,random_state = 123).fit(iris_data) 
    score = silhouette_score(iris_data,kmeans.labels_)
    silhouettteScore.append(score)
plt.figure(figsize=(10,6))
plt.plot(range(2,15),silhouettteScore,linewidth = 1.5,linestyle = '-')
plt.show()

使用Calinski-Harabasz指数评价K-Means聚类模型

from sklearn.metrics import calinski_harabaz_score
for i in range(2,7):
    kmeans = KMeans(n_clusters = i,random_state = 123).fit(iris_data) 
    score = calinski_harabaz_score(iris_data,kmeans.labels_)
    print('iris数据聚%d类calinski_harabaz指数为：%f'%(i,score))

iris数据聚2类calinski_harabaz指数为：513.303843
iris数据聚3类calinski_harabaz指数为：560.399924
iris数据聚4类calinski_harabaz指数为：529.120719
iris数据聚5类calinski_harabaz指数为：494.094382
iris数据聚6类calinski_harabaz指数为：474.753604