基于二乘法的数据预测算法

最新推荐文章于 2024-11-23 18:07:34 发布

原创最新推荐文章于 2024-11-23 18:07:34 发布 · 566 阅读

0 ·

CC 4.0 BY-SA版权

机器学习系列专栏收录该内容

2 篇文章

订阅专栏

本文介绍了一个基于线性回归的销量预测模型，该模型使用了最近8天的销售数据来预测接下来一天的数据。通过Python实现，利用JetBrains PyCharm 2017.2.3进行开发。

原理：根据最近8天的销售数据预测最近一天的销售数据

开发工具：JetBrains PyCharm 2017.2.3

代码：

from numpy import *

def standRegres(xArr,yArr):
    xMat = mat(xArr); yMat = mat(yArr).T
    xTx = xMat.T*xMat
    if linalg.det(xTx) == 0.0:
        print("This matrix is singular, cannot do inverse")
        return
    ws = xTx.I * (xMat.T*yMat)
    return ws

def loadDataSet(fileName):
    allDataMat = []
    dataMat = []
    yMat = []
    fr = open(fileName)
    for line in fr.readlines():
        curLine = line.strip().split('\t')
        allDataMat.append(float(curLine[0]))

    ii = 0
    for j in range(len(allDataMat)):
        if (j > 8):
            lineArr = []
            lineArr.append(allDataMat[j - 8])
            lineArr.append(allDataMat[j - 7])
            lineArr.append(allDataMat[j - 6])
            lineArr.append(allDataMat[j - 5])
            lineArr.append(allDataMat[j - 4])
            lineArr.append(allDataMat[j - 3])
            lineArr.append(allDataMat[j - 2])
            lineArr.append(allDataMat[j - 1])
            dataMat.append(lineArr)
            yMat.append(allDataMat[j])
    return yMat,dataMat

#加载数据--X为前8个连续数据。Y为当前值
yArr,xArr = loadDataSet('myDemo.txt')
#print(yArr)
#print(xArr)
#求以前8个数为参数的回归系统（8个常量）
ws = standRegres(xArr,yArr)
print(ws)
#print("3.176513	3.960008".split('\t'))

print('-------------------------')
xMat = mat(xArr)
yMat = mat(yArr)
yHat = xMat*ws

print(yMat.tolist())
print(yHat[:,0].flatten().A[0])
print('-------------------------')
#基于总方差求平均差（波动范围）
print(sqrt(sum((yArr-yHat.T.flatten().A[0])**2)/yHat.size))

'''
print(yArr)
print(corrcoef(yHat.T,yMat))
'''
#画图
import matplotlib.pyplot as plt
fig = plt.figure()
ax = fig.add_subplot(111)

a=[]
for i in range(len(xMat[:,1].flatten().A[0])):
    a.append(i)

ax.scatter(a,yMat.T[:,0].flatten().A[0])
ax.plot(a,yHat)
'''
xCopy = xMat.copy()
xCopy.sort(0)
yHat = xCopy*ws
ax.plot(xCopy[:,1],yHat)

'''
plt.show()

训练数据：

myDemo.txt

3.176513
3.816464
4.550095
4.256571
4.560815
3.929515
3.526170
3.156393
3.110301
3.149813
3.476346
4.119688
4.282233
3.486582
4.655492
3.965162
3.514900
3.125947
4.094115
3.476039
3.210610
3.190612
4.631504
4.295890
3.085028
3.448080
3.167440
3.364266
3.993482
3.891471
3.143259
3.114204
3.851484
4.621899
4.580768
3.620992
3.580501
4.618706
3.676867
4.641845
3.175939
4.264980
3.558448
3.436632
3.831052
3.182853
3.498906
3.946833
3.900583
4.238522
4.233080
3.521557
3.203344
4.278105
3.555705
3.502661
3.859776
4.275956
3.916191
3.587961
3.183004
4.225236
4.231083
4.240544
3.222372
4.021445
3.567479
3.562580
4.262059
3.208813
3.169825
4.193949
3.491678
4.533306
3.550108
4.636427
3.557078
3.552874
3.494159
3.206828
3.195266
4.221292
4.413372
4.184347
3.742878
3.201878
4.648964
3.510117
3.274434
3.579622
3.489244
4.237386
3.913749
3.228990
4.286286
4.628614
3.239536
4.457997
3.513384
3.729674
3.834274
3.811155
3.598316
4.692514
4.604859
3.864912
3.184236
3.500796
3.743365
3.622905
4.310796
3.583357
3.901852
3.233521
3.105266
3.865544
4.628625
4.231213
3.791149
3.968271
4.253910
3.194710
3.996503
3.904358
3.503976
4.557545
3.699876
4.613614
3.140401
4.206717
3.969524
4.476096
3.136528
4.279071
3.200603
3.299012
3.209873
3.632942
3.248361
3.995783
3.563262
3.649712
3.951845
3.145031
3.181577
4.637087
3.404964
3.873188
4.633648
3.154768
4.623637
3.078132
3.913596
3.221817
3.938071
3.880822
4.176436
4.648161
3.332312
4.240614
4.532224
4.557105
4.610072
4.636569
4.229813
3.500860
4.245514
4.605182
3.454340
3.180775
3.380820
4.565020
3.279973
4.554241
4.633520
4.281037
3.844426
3.891601
3.849728
3.492215
4.592374
4.632025
3.756750
3.133555
3.567919
4.363382
3.560165
4.564305
4.215055
4.174999
4.586640
3.960008
3.529963
4.213412
3.908685
3.585821
4.374394
3.213817
3.952681
3.129283