问题遇到的现象和发生背景
用代码块功能插入代码，请勿粘贴截图
import pandas as pd
import numpy as np
import matplotlib
matplotlib.use('TkAgg')
data_set = pd.read_csv('C:\\Users\\pc.000\\Desktop\\数据集\\iris.csv', delimiter=',', header=None)
x_train = data_set.iloc[0:100, 0:4].values.T
x_test = data_set.iloc[100:151, 0:4].values.T
# y_train = Train_Y[0:100]
# y_test = Train_Y[100:151]
def sigmiod(x):
return 1/(1+np.exp(-x))
def dsigmiod(y):
return y*(1-y)
n, m = np.shape(x_train)
n_x = 4
n_h = 2
n_y = 4
np.random.seed(2)
w1 = np.random.randn(n_h, n_x)*0.1
b1 = np.zeros((n_h, 1))
w2 = np.random.randn(n_y, n_h)*0.1
b2 = np.zeros((n_y, 1))
def forward(w1, x_train, b1, w2, b2):
z1 = np.dot(w1, x_train) + b1
a1 = sigmiod(z1)
z2 = np.dot(w2, a1) + b2
a2 = z2
return z1, a1, z2, a2
def costfuction(a2, x_train):
error =np.sum(0.5*(a2 - x_train)**2)/m
return error
def backward(y_train, a2, a1, w2, x_train):
dz2 = a2 - y_train # 1 * 90
dw2 = np.dot(dz2, a1.T) # 1 * 90 * 90 * 5 = 1 * 5
db2 = np.sum(dz2, axis=1, keepdims=True)/m
dz1 = np.dot(w2.T, dz2) * dsigmiod(a1)
dw1 = np.dot(dz1, x_train.T)
db1 = np.sum(dz1, axis=1, keepdims=True)/m
return dz2, dw2, db2, dz1, dw1, db1
alpha = 0.0001
number = 15000
for i in range(1, number+1):
z1, a1, z2, a2 = forward(w1, x_train, b1, w2, b2)
error = np.sum(0.5*(a2 - x_train)**2)/m
dz2, dw2, db2, dz1, dw1, db1 = backward(x_train, a2, a1, w2, x_train)
w1 = w1 - alpha * dw1
w2 = w2 - alpha * dw2
b1 = b1 - alpha * db1
b2 = b2 - alpha * db2
if i % 1000 == 0:
print("迭代次数", i)
print("误差为", error)
z1_test, a1_test, z2_test, a2_test = forward(w1, x_test, b1, w2, b2)
n_test, m_test = np.shape(x_test)
error_test = np.sum(0.5 * (a2_test - x_test) ** 2) / m_test
print("测试的误差为", error_test)