Machine Learning 第五波编程作业 – Regularized Linear Regression and Bias/Variance
仅列出核心代码：
1.linearRegCostFunction.m
h = X * theta;
J = (X * theta - y).' * (X * theta - y) / (2*m)...
+(lambda/(2*m)) * sum(theta(2:end).^2);
grad = grad(:);
grad(1) = (X(:, 1).' * (h - y)) /m;grad(2:end) = (X(:, 2:end).' * (h - y)) /m ...
+ (lambda/m) * theta(2:end);
2.learningCurve.m
for i = 1:m
Xi = X(1:i, :);
yi = y(1:i);
lambda = 1;
[theta] = trainLinearReg(Xi, yi, lambda);
lambda = 0;
% For train error, make sure you compute it on the training subset
[error_train(i), ~] = linearRegCostFunction(Xi, yi, theta, lambda);
% For validation error, compute it over the entire cross validation set
[error_val(i), ~] = linearRegCostFunction(Xval, yval, theta, lambda);
end
3.polyFeatures.m
for i =1:p
X_poly(:, i) = X(:, 1).^i;
end
4.validationCurve.m
for i = 1:length(lambda_vec)
[theta] = trainLinearReg(X, y, lambda_vec(i));
% For train error, make sure you compute it on the training subset
[error_train(i), ~] = linearRegCostFunction(X, y, theta, 0);
% For validation error, compute it over the entire cross validation set
[error_val(i), ~] = linearRegCostFunction(Xval, yval, theta, 0);
end
