Python—对乳腺癌数据集预处理和分类

题目：

已知 UCI 数据集 breast-cancer-wisconsin，breast-cancer-wisconsin 是肿瘤学家研究切片组织，描述组织各种特征决定肿瘤是良性还是恶性的数据集，数据集共有699个样本个数，有11个特征，第一个为id number，最后一个为class(有无癌症的分类)，该数据集包含若干个缺失数据。要求：
(1)首先对缺失数据进行处理，并说明处理的方法。
(2)随机选取数据集的 70%的数据构成训练集，剩余30%数据构成测试集，并应用逻辑回归算法对测试集进行分类，采用Accuracy作为评估算法的标准。
(3)在(1)(2)的基础上采用5-折交叉验证方式进行试验，得到实验的Accuracy值。

步骤：

# -*- coding: utf-8 -*-
#####一：作业提交#####
import pandas as pd 
# from scipy.interpolate import lagrange

inputfile='C:/Users/PEXYGGJF/Desktop/ly/text/breast-cancer-wisconsin.csv'
outputfile='C:/Users/PEXYGGJF/Desktop/ly/work/ly.csv'

data=pd.read_excel(inputfile)
data['Bare Nuclei'][(data['Bare Nuclei']<0) | (data['Bare Nuclei']>11)]=None


import pandas as pd
import numpy as np
 
column_names = ['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 = pd.read_csv('https://archive.ics.uci.edu/ml/machine-learning-databases/breast-cancer-wisconsin//breast-cancer-wisconsin.data',names=column_names)

data = data.replace(to_replace='?',value=np.nan)    #非法字符的替代，缺失值处理
data = data.dropna(how='any')        #去掉空值，any：出现空值行则删除
print(data.shape)
print(data.head())

##################################################################################
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import classification_report
# from sklearn.externals 
import joblib

import pandas as pd
import numpy as np



def logistic():
    """
    逻辑回归做二分类进行癌症预测（根据细胞的属性特征）
    
    """
    # 构造列标签名字,一共11个
    column = ['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']

    # 读取数据
    data = pd.read_csv("E:/文档资料/python数据分析和数据挖掘/work/python report/breast-cancer-wisconsin.csv",
                       engine='python',names=column)

    print(data)

    # 缺失值进行处理
    data = data.replace(to_replace='?', value=np.nan)

    data = data.dropna()

    # 进行数据的分割，1到10列为特征值，11列为目标值
    x_train, x_test, y_train, y_test = train_test_split(data[column[1:10]], data[column[10]], test_size=0.25)

    # 进行标准化处理
    std = StandardScaler()

    x_train = std.fit_transform(x_train)
    x_test = std.transform(x_test)

    # 逻辑回归预测
    lg = LogisticRegression(C=1.0)

    lg.fit(x_train, y_train)

    print(lg.coef_)

    y_predict = lg.predict(x_test)

    print("准确率：", lg.score(x_test, y_test))

    print("召回率：", classification_report(y_test, y_predict, labels=[2, 4], target_names=["良性", "恶性"]))

    #保存训练好的模型
    joblib.dump(lg, "./lg.pkl")
    
    #加载模型，预测自己的数据
    model = joblib.load("./lg.pkl")
    
    # 读取数据，数据为相同格式下需要预测的数据。和前文七、（1）（2）的操作一样
    data = pd.read_csv("E:/文档资料/python数据分析和数据挖掘/work/python report/new-breast-cancer-wisconsin.csv",
                       engine='python',names=column)

    xx_test= data[column[1:10]]  #获取特征值1 到 10 列，
    xx_test = std.transform(xx_test)
    
    yy_predict = model.predict(xx_test)

    print("保存的模型预测的结果：", yy_predict)

if __name__ == "__main__":
    logistic()
    

# -*- coding:utf-8 -*-
#####二：较标准呈现#####
import numpy as np   #导入numpy库
import random        #导入random库，产生随机数
import csv           #导入csv格式文件

def loadDataSet():            #加载数据，标签，特征值
    trainMat = [];data0=[]    #训练集数组，测试集数组
    data = csv.reader(open('E:/pywork/test/sy-3/breast-cancer-wisconsin.csv'))
    for line in data:         #读入每一行数据，判断数据集中是否有“？”，对缺失值进行处理
        if "?" in line:
            continue
        lineArr = []
        if line[0] != '':     
            for i in range(2, 12):                       #读取第二列到第十一列的标签的每一行数据
                lineArr.append(float(line[i])) 
            data0.append(lineArr)                        #添加数据
    m ,n = np.shape(data0)
    times = int(m*0.7)                                   #70%分割线
    for i in range(times):
        randIndex = int(random.uniform(0,len(data0)))    #产生随机数
        trainMat.append(data0[randIndex])
        del(data0[randIndex])
    testMat = data0[:]
    return trainMat, testMat                             #返回测试集，训练集

def depart(dataset):
    dataMat =[] ; labelMat = []
    for line in dataset:
        lineArr = []
        for i in range(9):
            lineArr.append(line[i])
        dataMat.append(lineArr)  
        labelMat.append(line[9])
    return  dataMat,labelMat

def sigmoid(inX):                                                #S函数
    return 1.0/(1+np.exp(-inX))

def stocGradAscent1(dataMatrix,LabelMat,numIter=500):            #随机梯度上升算法
    m,n = np.shape(dataMatrix)
    weights = np.ones(n)
    for j in range(numIter):
        dataIndex =list( range(m))
        for i in range(m):      
            alpha = 4/(1.0+j+i)+0.01                             #每次迭代时更新alpha值
            randIndex = int(random.uniform(0,len(dataIndex)))    #随机选取更新
            h = sigmoid(sum(dataMatrix[randIndex]*weights))
            error = LabelMat[randIndex] - h
            weights = weights + alpha * error * dataMatrix[randIndex]
            del(dataIndex[randIndex])
    return weights

def classifyVector(inX,weights):                                 #逻辑回归分类函数，inX特征向量，weights回归系数
    prob = sigmoid(sum(inX*weights))  
    if prob > 0.5:return 2.0
    else:return 4.0

def colicTest():                                                 #打开测试集和训练集，并对数据进行格式化处理
    trainMat, testMat = loadDataSet()
    trainSet, trainLabel = depart(trainMat)                      #导入类别标签为最后一项
    trainWeights = stocGradAscent1(np.array(trainSet),trainLabel,500)              #使用随机梯度上升算法计算回归系数向量
    rightCount = 0; numTestVec = 0.0
    for line in testMat:                                         #格式化测试集
        numTestVec += 1.0
        lineArr = []
        for i in range(9):                                       #导入特征值，有9个特征
            lineArr.append(float(line[i]))
        if int(classifyVector(np.array(lineArr),trainWeights)) == int(line[9]):    #使用训练集计算出回归系数对测试集进行分类，并对比测试集的类别标签，计算错误数量
            rightCount += 1
    rightCount = (float(rightCount)/numTestVec)                  #正确率
    print ("The right rate of this test is: %f" % rightCount)
    return rightCount


def multiTest():                                                 #colicTest函数10次，取正确率平均值
    numTests = 10; rightSum = 0.0
    for k in range(numTests):
        rightSum += colicTest()
    print ("after %d iterations the average error rate is: %f" % (numTests, rightSum/float(numTests)))


multiTest()