EP1.numpy实现感知机

import numpy as np

def sigmoid(x):
  return 1/(1+np.exp(-x)) 
  
def train():
#设定参数值
    m=train_x.shape[1]
    learning_rate=0.01
    iteration_nums=10000
    input_dim=train_x.shape[0]
#初始化参数    
    b=0
    w=np.zeros(shape=(input_dim,1))
    for i in range(iteration_nums):
#前向计算损失(交叉熵函数)
        res=sigmoid(np.dot(w.T,train_x)+b)
        cost=-1/m*np.sum(train_y*np.log(res)+(1-train_y)*np.log(1-res))
        print("cost after iteration %i:%f" % (i, cost))
#后向迭代更新参数
        dw=np.dot((res-train_y).T,train_x)/m
        db=np.sum(res-train_y)/m
        w=w-dw*learning_rate
        b=b-db*learning_rate
    return w,b

def predict():
    train()
    test_y=sigmoid(np.dot(w.T,test_x)+b)
    return test_y