机器学习基石作业三20题

import numpy as np
import requests

def get_data(url):
    content=requests.get(url).content
    content=content.decode('utf-8')
    content=content.split("\n")
    X=[]
    Y=[]
    for line in content[:-1]:
        data=line.split()
        y=float(data[-1])
        Y.append(y)
        x=data[:-1]
        for i in range(len(x)):
            x[i]=float(x[i])
        X.append([1.0]+x)
    X=np.mat(X)
    Y=np.array(Y)
    return X,Y

def init(X):
    return np.zeros((X.shape[1],1))

def sigmoid(s):
    return 1/(1+np.exp(-s))


def grad(w,X,Y,i):
    s=-Y[i]*X[i]*w
    return -Y[i]*sigmoid(s).item()*X[i].T

def sign(v):
    if v>=0.5:
        return 1
    else:
        return -1
def cost(w,X,Y):
    err=0
    y_hat=sigmoid(X*w).T
    y_hat=y_hat.tolist()
    for i in range(len(Y)):
        if(sign(y_hat[0][i])!=Y[i]):
            err+=1
    return err/len(Y)

if __name__=="__main__":
    trainX,trainY=get_data("https://www.csie.ntu.edu.tw/~htlin/mooc/datasets/mlfound_algo/hw3_train.dat")
    testX,testY=get_data("https://www.csie.ntu.edu.tw/~htlin/mooc/datasets/mlfound_algo/hw3_test.dat")
    w=init(trainX)
    l_r=0.01
    for i in range(2000):
        w=w-l_r*grad(w,trainX,trainY,i%len(trainY))
    print(cost(w,testX,testY))