我在学习机器学习课程，并尝试在Matlab中实现梯度下降算法。单独测试时，computeCost函数运行正常。我用它来查看每次迭代的成本，但似乎成本根本没有减少。成本只是随机波动。alpha的值被给定为0.01，所以我知道这不是学习率过大的问题。我得到的theta值与预期输出相差很大。我哪里做错了？提前谢谢！

function theta = gradientDescent(X, y, theta, alpha, num_iters)%GRADIENTDESCENT Performs gradient descent to learn theta% Initialize some useful valuesm = length(y); % number of training examplestemp1=0;temp2=0;for iter = 1:num_itersfor k = 1:m    temp1 = temp1 + (theta(1) + theta(2)*X(k, 2) - y(k));    temp2 = temp2 + ((theta(1) + theta(2)*X(k, 2) - y(k))*X(k, 2));end theta(1) = theta(1)-(1/m)*alpha*temp1;theta(2) = theta(2)-(1/m)*alpha*temp2;computeCost(X, y, theta)endend

编辑：这里也附上computeCost函数

function J = computeCost(X, y, theta)m = length(y); % number of training examplesJ = 0;temp = 0;for index = 1:m    temp = temp + (theta(1) + theta(2)*X(index, 2)-y(index))^2;end J = temp/(2*m); end

回答：

尝试更改：

temp1=0;temp2=0;for iter = 1:num_iters

为

for iter = 1:num_iters  temp1=0;  temp2=0;

梯度需要在每次迭代时重新计算（否则你实际上是在引入动量项）。