Regularized linear regression gradient

ex5/linearRegCostFunction.m

@@ -22,6 +22,8 @@ grad = zeros(size(theta));
 J =   1/(2*m) * sum(((X*theta)-y).^2) ...
     + lambda/(2*m) * sum(theta(2:end).^2);
 
+grad = 1/m * X' * ((X*theta)-y) + (lambda/m) * [0; theta(2:end)];
+
 % =========================================================================
 
 grad = grad(:);