Pytorch实战之线性回归代码实现

思维导图

main.py文件代码如下：

import torch #导入torch包
import numpy as np
# y=wx+b
def comput_e_error_for_line_given_points(b, w, points):
    totalEorr = 0
    for i in range(0,len(points)):
        x=points[i,0]
        y=points[i,1]
        totalEorr+=(y-(w*x+b))**2
    return totalEorr/float(len(points))

def step_gradient(b_current,w_current,points,learningRate):
    b_gradient=0
    w_gradient=0
    N=float(len(points))
    for i in range(0,len(points)):
        x=points[i,0]
        y=points[i,1]
        b_gradient += -(2/N)*(y-((w_current*x)+b_current))
        w_gradient += -(2/N)*x*(y-((w_current*x)+b_current))
    new_b=b_current-(learningRate*b_gradient)
    new_w=w_current-(learningRate*w_gradient)
    return [new_b,new_w]

def gradient_descent_runner(points,starting_b,starting_w,learning_rate,num_interation):
    b = starting_b
    w =starting_w
    for i in range(num_interation):
        b,w =step_gradient(b,w,np.array(points),learning_rate)
    return [b,w]

def run():
    points =np.genfromtxt("data.csv", delimiter=",")
    leaning_rate=0.001
    initial_b=2
    initial_w=3
    num_iteration=10000
    print("Starting gradient descent at b ={0},m={1},error ={2}"
          .format(initial_b,initial_w
                      ,comput_e_error_for_line_given_points(initial_b,initial_w,points))
          )
    print("Running")
    [b,w] = gradient_descent_runner(points,initial_b,initial_w,leaning_rate,num_iteration)
    print("After{0} iteations b ={1} w={2} error={3}".format(num_iteration,b,w,comput_e_error_for_line_given_points(b,w,points))
          )
run()

生成数据集代码如下(data.csv) ：

import numpy as np
from matplotlib import pyplot as plt

# 数据集的要求如下，
# 训练集 xdata为100个在（0,1）范围内服从正态分布的随机数，
#       ydata满足：ydata=2*xdata+3+random1,其中random1也是（0,1）范围内服从正态分布的随机数。
# 测试集：选择训练集中前30个数据作为测试集

# 生成训练集xdata
for i in range(0,100):
    xdata = np.random.random()
   # ydata[i] = np.zeros((100,1))
    random1 = np.random.random()
    ydata = 2 * xdata+3+random1
    print("{0},{1}".format(xdata, ydata))
    

运行效果展示如下：

线性回归数据集样例(data.csv）：

0.7959268131949154,5.383253081048107
0.10792226843451991,3.5672430653843614
0.2236481182771649,4.382010400684081
0.891695496410794,5.455785790798779
0.4107177706568511,3.9167518925643465
0.7098110526762221,5.199094617324766
0.5726718092577726,4.386343890884308
0.850689677164127,4.960650345986238
0.527341989113963,4.418461070582508
0.5407383809937704,4.5627824491847955
0.9393720199081832,5.407068167027157
0.404111321556505,3.903080094784834
0.5876819809625564,4.421394423335027
0.7566599371624856,5.240679612703435
0.0031178329236888347,3.8442504171697847
0.270310921739623,3.983739867944232
0.6643683517996607,4.359459463576226
0.9527310211083514,4.968393331955517
0.2892343760050178,4.040681819521732
0.46474848712503447,4.013226297178308
0.38899395293092753,3.7954249709832015
0.8393131259276586,5.601939985957008
0.10964512586659714,3.94851985635711
0.6261757159186817,4.379742390947595
0.08565667598505655,3.606665516431419
0.2370180363736345,3.9625511283429713
0.7230523363252718,4.7348974176283365
0.7197066759784131,4.678245747524764
0.8806035990816686,5.193817064942353
0.876398017062797,5.600633869952809
0.8933393915717712,5.575496230757278
0.7068823078576177,4.778773075393766
0.709033905181841,4.732462037265902
0.8402420992227093,4.695932293253544
0.10837407805690058,3.2415365443499438
0.5631490491018708,4.373276665217919
0.4472270081739842,4.548709891719379
0.19150799074937952,3.4555041178151438
0.7821204649835846,4.833793474769088
0.5447451854388814,4.424243330219422
0.5536264693840504,4.907500684923544
0.8032002914486217,4.827199696053637
0.4244326395755429,4.540430034046199
0.6571746863360602,4.916846569671098
0.39928751719262146,3.9128596629191783
0.8461790206375631,4.715795724237515
0.48164498728976635,4.0434483299889585
0.21243940642655335,3.9350069589576737
0.5459791966500642,4.178299647642944
0.15591629833189136,4.208862653825564
0.3318545713261687,4.607069147207268
0.31337560872349335,3.681321872351009
0.8746815858444879,5.272427255091438
0.39391686710395657,4.388534391646889
0.4793015019357064,3.967322943039812
0.050359282562026375,3.15624439605456
0.4346404837996358,4.824577504350639
0.5637581740174465,4.498518989232363
0.7234594570524097,4.829171724018326
0.6588940040417601,5.0248055753643115
0.10455070414976741,3.287324564852888
0.8472013204912199,5.220390502713326
0.8993465276185247,5.292857170440293
0.25408619265199506,4.456075856474079
0.33545554644583775,4.135687745216281
0.531801937690468,4.618679650306403
0.6967017410342263,5.159739489522426
0.9323388229203458,5.0692772708558405
0.14683451371051082,3.384713814368769
0.10370090011007094,3.6740992437380076
0.9527298299437502,5.354806096815416
0.09349077824028018,3.38987042788691
0.12945204991905068,3.966301687383356
0.2935058975079232,3.8480254159268306
0.5972685377661336,4.532457919626502
0.5307169141975102,4.826574087837489
0.4265298582030739,4.50333560555219
0.27716945105123814,4.2856318500195645
0.7218085392053041,4.749430370829406
0.3926147186480382,4.027877145111341
0.860373996404969,5.146214100068526
0.9164588262130305,5.57223206392356
0.6491618151006234,4.663168027847816
0.15535218391420036,3.9910332036211984
0.3725747901005698,4.250937515531864
0.4704407827973769,4.699327725407754
0.1165160789572447,3.9365087193959853
0.8392828002905782,5.1504058711641365
0.9305998527731457,5.461888219112742
0.5657952945675231,4.549410660145753
0.8936753318690827,5.633624420799182
0.30145455997523884,3.698892542353196
0.017762394148341465,3.800492967019841
0.4850334935570817,4.95502562398627
0.9161554507745315,5.4506454086330685
0.5105198903372674,4.480026680840481
0.06874387575796892,3.318011713652211
0.8278604344296541,5.264489583300447
0.9003304092956863,5.048744821834357
0.9824348313198714,5.636684504177316