多元线性回归

#DATASET: https://archive.ics.uci.edu/ml/datasets/Computer+Hardware

import numpy as np
import matplotlib.pyplot as plt
import pandas as pd
import sklearn
import sklearn.preprocessing as pre

df=pd.read_csv('datamachineimachine.data')
df.head()

	vendor	name	MYCT	MMIN	MMAX	CACH	CHMIN	CHMAX	PRP	ERP
0	adviser	32/60	125	256	6000	256	16	128	198	199
1	amdahl	470v/7	29	8000	32000	32	8	32	269	253
2	amdahl	470v/7a	29	8000	32000	32	8	32	220	253
3	amdahl	470v/7b	29	8000	32000	32	8	32	172	253
4	amdahl	470v/7c	29	8000	16000	32	8	16	132	132

#convert string to num
#X includes the first 8 attributes, Y is the target prediction

x1=pre.LabelEncoder().fit_transform(df.iloc[:,0])
x2=pre.LabelEncoder().fit_transform(df.iloc[:,1])
X=np.zeros((df.iloc[:,:-2].shape[0],8))
X[:,0]=x1
X[:,1]=x2
X[:,2:]=df.iloc[:,2:-2]
Y=df.iloc[:,-2] #goal

#convert string to num
#X includes the first 8 attributes, Y is the target prediction

x1=pre.LabelEncoder().fit_transform(df.iloc[:,0])
x2=pre.LabelEncoder().fit_transform(df.iloc[:,1])
X=np.zeros((df.iloc[:,:-2].shape[0],8))
X[:,0]=x1
X[:,1]=x2
X[:,2:]=df.iloc[:,2:-2]
Y=df.iloc[:,-2] #goal

#compute loss, train parameters

count=0
while True:
    count+=1
    predict_y=np.dot(X,W.T)+b
    rloss=np.dot((Y-predict_y).T,Y-predict_y)/(2*m)
    w_gradient=np.dot((Y-predict_y).T,X)*(-1.0/m)
    b_gradient=np.dot((Y-predict_y).T,np.ones(shape=[m,]))*(-1.0/m)
    #gradient descent
    W-=w_gradient*alpha
    b-=b_gradient*alpha
    if count%10000==0:
        print(rloss)
    if rloss<2500:
        print('MSEloss={}'.format(rloss))
        break

MSEloss=2456.133718645051

W

array([-0.02520069, -0.17961688, -0.00338619,  0.01356295,  0.00726707,
        0.16543731,  0.01494539,  0.13264917])