逻辑回归乳腺癌预测

根据历史女性乳腺癌患者数据集（医学指标）构建逻辑回归分类模型进行良／恶性乳腺癌肿瘤预测
详细数据见社区下载https://www.kesci.com/home/project/5eef43e2caa99b002d6e7a16
也可以在UCI机器学习网站下载

#-*- coding: utf-8 -*-
"""created on 2020/6/21
@author: wangshisuifeng
"""

from sklearn.linear_model import LogisticRegression
import pandas as pd
import numpy as np
from sklearn.metrics import mean_squared_error,classification_report
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler

def log():
    #加载数据集
    data = pd.read_csv('E:\\Desktop\\breast-cancer-wisconsin1.csv')
    #进行数据处理
    data = data.replace(to_replace='?',value=np.nan)
    data = data.dropna()
    x_train = data[['number','height','size','avg','li','once_size','he','ransezhi','heren','yousifenlie']]
    y_train = data['class']
    #分隔训练集和测试集
    x_train,x_test,y_train,y_test = train_test_split(x_train,y_train,train_size=0.25)
    #标准化数据
    std_y = StandardScaler()
    x_train = std_y.fit_transform(x_train)
    x_test = std_y.transform(x_test)
    #利用逻辑回归预测
    logi = LogisticRegression(C=1.0)
    logi.fit(x_train,y_train)
    print('逻辑回归系数:',np.ravel(logi.coef_))
    y_logi_pre = logi.predict(x_test)
    print(y_logi_pre)
    #准确率
    print('精确率',logi.score(x_test,y_test))
    #召回率
    print('召回率',classification_report(y_test,y_logi_pre,labels=[2,4],target_names=['良性','恶性']))
    return None
if __name__=='__main__':
    log()