'''
1、经过多次迭代计算出最佳的回归系数，在每次迭代中增加或减少某个权重观察其错误率变化。选择对误差影响小的方向：

'''
from numpy import *

#import numpy as np
import pandas as pd

def rssError(yArr, yHatArr):
    return ((yArr - yHatArr) ** 2).sum()

###2、加载数据 ，txt格式以\t分割  将文件数据转换成列表返回dataMat, labelMat
def loadDataSet(fileName):
    numFeat = pd.read_csv(fileName,index_col=0)
    dataMat = array(numFeat.iloc[:500, 2:]).tolist()
    labelMat = array(numFeat.iloc[:500, 1]).tolist()
    #print(len(dataMat[0]),len(dataMat),dataMat[:5],type(dataMat),labelMat[:10])
    return dataMat, labelMat
# fileName = 'edu_shaoer_afterfeature.csv'
# loadDataSet(fileName)
# def loadDataSet(fileName):
#     numFeat = len(open(fileName).readline().split('\t')) - 1
#     dataMat = []
#     labelMat = []
#     fr = open(fileName)
#     for line in fr.readlines():
#         lineArr = []
#         curLine = line.strip().split('\t')
#         for i in range(numFeat):
#             lineArr.append(float(curLine[i]))
#         dataMat.append(lineArr)
#         labelMat.append(float(curLine[-1]))
#     return dataMat, labelMat


###3、标准化数据
def regularize(xMat):  # regularize by columns
    inMat = xMat.copy()
    inMeans = mean(inMat, 0)  # 按列取均值
    inVar = var(inMat, 0)  # 按列取方差
    inMat = (inMat - inMeans) / inVar
    #print(type(inMat),inVar)
    return inMat


###4、
def stageWise(xArr, yArr, eps=0.01, numIt=100):
    xMat = mat(xArr)  # 将xArr变成矩阵形式
    yMat = mat(yArr).T  # 将yArr变成矩阵形式->m X 1
    yMean = mean(yMat, 0)  # 将yMat按列求均值
    yMat = yMat - yMean
    print(yMat[:10])
    xMat = regularize(xMat)

    m, n = shape(xMat)
    returnMat = zeros((numIt, n))  # 初始化一个100 X n的全0矩阵
    ws = zeros((n, 1))  # 初始化一个n X 1的全0矩阵
    wsTest = ws.copy()
    wsMax = ws.copy()

    for i in range(numIt):  # numIt:迭代的次数
        #print(ws.T)  # 输出一个 1 X n的全0矩阵
        lowestError = inf  # 误差初始化为正无穷
        for j in range(n):  # 对每个特征
            for sign in [-1, 1]:  # 分别计算增加或减少该特征对误差的影响
                wsTest[j] += eps * sign  # eps:每次迭代移动的步长
                yTest = xMat * wsTest
                rssE = rssError(yMat.A, yTest.A)  # 得到平方误差进行比较; 矩阵名.A: 矩阵转化为array数组类型
                if rssE < lowestError:
                    lowestError = rssE
                    wsMax = wsTest
                    #print('wsmax',wsMax)
        ws = wsMax.copy()
        returnMat[i, :] = ws.T
    print(returnMat[:5],returnMat.shape,)
    return returnMat  # 迭代结束返回回归系数


def main():
    xArr,yArr = loadDataSet('edu_shaoer_afterfeature.csv')
    stageWise(xArr,yArr,0.1,5000)
if __name__ == '__main__':
    main()

#——————————————————————————————————————————
#岭回归

### Ridge Regression ###
def ridgeRegres(xMat, yMat, lam=0.2):
    xTx = xMat.T * xMat
    denom = xTx + lam*eye(shape(xMat)[1])
    if linalg.det(denom) == 0.0:
        print ("This matrix is singular, cannot do inverse")
        return
    ws = denom.I * (xMat.T*yMat)
    return ws

def ridgeTest(xArr, yArr):
    xMat = mat(xArr)
    xMeans = mean(xMat, 0)
    xVar = var(xMat, 0)
    xMat = (xMat-xMeans) / xVar
    yMat = mat(yArr).T
    yMean = mean(yMat, 0)
    yMat = yMat - yMean
    numTestPts = 30
    wMat = zeros((numTestPts, shape(xMat)[1]))
    for ii in range(numTestPts):
        ws = ridgeRegres(xMat, yMat, exp(ii-10))
        wMat[ii,:] = ws.T
    return wMat