我正在尝试用Python而不是matlab解决Andrew NG的第3周课程分配问题,我发现fmin_bfgs与ex2.m文件中的fminunc类似。
问题是,当我尝试输入cost函数和将梯度计算为fmin_bfgs的函数时,它返回一个回溯,内容为: TypeError:“ numpy.ndarray”对象不可调用。 我在尝试调试问题时遇到了困难,因此将不胜感激
# code that produces the error:
results = opt.fmin_bfgs(costFunction.cost_function(weights, data_input, desired_output), weights,
fprime=costFunction.compute_grad(weights, data_input, desired_output), maxiter=400)
#the costFunction module:
import numpy as np
import sigmoid
def cost_function(theta, x, y):
size = np.shape(x)
z = np.dot(x, theta)
h = sigmoid.sigmoid(z)
j = (1 / size[0]) * (np.dot(-y, np.log(h)) - np.dot((1 - y), np.log(1 - h)))
return j
def compute_grad(theta, x, y):
size = np.shape(x)
z = np.dot(x, theta)
h = sigmoid.sigmoid(z)
error = np.subtract(h, y)
grad = (1 / size[0]) * np.dot(np.transpose(error), x)
return grad