基于逻辑回归的癌症分析

import numpy
import pandas
from  sklearn.metrics import  roc_auc_score
from  sklearn.preprocessing import  StandardScaler  # 标准化
from sklearn.linear_model import LogisticRegression #逻辑回归
from  sklearn.metrics import  classification_report #召回率
from sklearn.model_selection import train_test_split  # 训练集 测试集 拆分

# 加载数据
data = pandas.read_csv('data/breast-cancer-wisconsin.data',header=None)

# 列名称
columns = ['Sample code number','Clump Thickness','Uniformity of Cell Size',
           'Uniformity of Cell Shape','Marginal Adhesion','Single Epithelial Cell Size',
           'Bare Nuclei','Bland Chromatin','Normal Nucleoli','Mitoses','Class']

data.columns = columns

print(data)

# 缺失值处理
data.replace('?',numpy.nan,inplace=True)

#检测缺失值
res_null = data.isnull().sum()


# 删除缺失值
data.dropna(axis=0,how='any',inplace=True)


# 进行筛选数据集 --- 去除编号的第一列
data = data.iloc[:,1:]


# 获取特征值
feature = data.iloc[:,:-1].values
target = data.iloc[:,-1].values

# 拆分数据
x_train,x_test,y_train,y_test = train_test_split(feature,target,test_size=0.3,random_state=1)


# 特征值需要标准化，目标值不需要标准化
stand = StandardScaler()

# 训练集的特征值
x_train  = stand.fit_transform(x_train)


# 测试集的特征值
x_test  = stand.fit_transform(x_test)

# 利用逻辑回归进行分类
lr = LogisticRegression() # 参考sgd下降优化
# 训练数据
lr.fit(x_train,y_train)

#进行预测数据
y_predict = lr.predict(x_test)

# 准确率
score = lr.score(x_test,y_test)

# 获取权重与偏置
weight = lr.coef_
bias = lr.intercept_

# 计算召回率
# 召回率越高越好--查的全
# f1-score -越高越好---体现模型越稳健
res_report = classification_report(y_test,y_predict,labels=[2,4],target_names=["良性","恶性"])

print(res_report)

# 如果 y_test>3设置为1，否则为0
y_test = numpy.where(y_test > 3,1,0)
print(y_test)

#计算AUC指标 ---针对的样本不平衡的状态
auc = roc_auc_score(y_test,y_predict)
print(auc)