机器学习-线性回归模型-西瓜书代码(LR)-RMSE

以下代码是本人在学习西瓜书时花费两个礼拜根据原理进行原创，若需转载请咨询本人，谢谢！

自我研究模拟

附上连续数值截图

数据截图：

运行截图：

Linear_config.py

"""
 Filename: Linear_config
 Author: kdd_zyx
 Description: 机器学习 - 线性回归
 Datas：kdd - 回归
 Start: 2018.11.3
 End: 2018.11.3
"""

import time as t
import numpy as np

start = 1   # 数据上标
end = -1    # 结尾下标

# str->float
def flt(pro_data):
    flt_data = [1.0]
    for data in pro_data:
        flt_data.append(float(data))

    return flt_data

def writ(NUM, result):
    try:
        with open('kdd{}.txt'.format(NUM), 'a+', encoding='UTF-8') as f:
            f.write(result)
            f.close()
    except Exception as e:
        print('Error：', e)
        writ(NUM, result)

# 引入训练集
def Lead_dataset():
    try:
        pro_dataset = []
        end_dataset = []
        txt_dataset = []
        f = open('train.txt', 'r', encoding='UTF-8')
        for line in f.readlines():
            data = line.strip().split(',')
            pro_dataset.append(flt(data[ :end]))
            end_dataset.append(float(data[end]))

        f = open('text.txt', 'r', encoding='UTF-8')
        for line in f.readlines():
            data = line.strip().split(',')
            txt_dataset.append(flt(data))


        return pro_dataset, end_dataset, txt_dataset

    except Exception as e:
        print('Error：', e)
        Lead_dataset(text_addr, text_splt)

Linear.py

"""
 Filename: Linear
 Author: kdd_zyx
 Description: 机器学习 - 线性回归
 Datas：kdd - 回归
 Start: 2018.11.3
 End: 2018.11.3
"""

from Linear_config import *

# 优化数据存储结构
class dataset:
    def __init__(self, pro_dataset, end_dataset):
        self.X = np.mat(pro_dataset)    # 特征数据集
        self.Y = np.mat(end_dataset).T  # 类别数据集
        self.m = np.shape(self.Y)[0]    # 数据行数
        self.W = np.mat(np.zeros((self.m, 1))) # 方程参数集

class Linear:
    def __init__(self, pro_dataset, end_dataset, txt_dataset):
        self.oS = dataset(pro_dataset, end_dataset)
        self.xT = np.mat(txt_dataset)

    def Verify(self):
        RMSE = 0

        for i in range(self.oS.m):
            yi = np.dot(self.oS.X[i, :], self.oS.W)
            print(str(i) + '\t\t' + str(round(float(yi), 5)) + '\t\t' + str(round(float(self.oS.Y[i]), 5)))
            RMSE += (yi - self.oS.Y[i]) ** 2

        Score = self.Cacl_Score(RMSE)

        try:
            for i in range(self.xT.shape[0]):
                yi = np.dot(self.xT[i, :], self.oS.W)
                # writ(0, str(float(yi)) + '\n')

        except Exception:
            pass

        finally:
            print("The rmse is:", round(float(Score), 3))  # 打印出正确率

    def Linear_Creat(self):
        Is_full = self.Is_full()

        if Is_full == 1:
            self.Full_mat()
            self.Verify()
        else:
            self.Not_full_mat()
            self.Verify()

    # 判断是否为 满秩矩阵 or 正定矩阵
    def Is_full(self):

        return 1

    def Linear_W(self):
        pass

    def Full_mat(self):
        self.oS.W = np.dot(np.dot(self.oS.X.T, self.oS.X).I, np.dot(self.oS.X.T, self.oS.Y))

    def Not_full_mat(self):
        pass

    def Cacl_Score(self, RMSE):

        RMSE /= self.oS.m
        RMSE = np.sqrt(RMSE)

        Score = 1 / (1 + RMSE)

        return Score

if __name__ == '__main__':

    sta_time = t.time()

    pro_dataset, end_dataset, txt_dataset = Lead_dataset()
    linear = Linear(pro_dataset, end_dataset, txt_dataset)
    linear.Linear_Creat()

    end_time = t.time()
    print("Time:", end_time - sta_time)