Mercurial > hg > machine-learning-hw3
changeset 5:cdab664cf253 default tip
Minor tweaks
| author | Jordi Gutiérrez Hermoso <jordigh@octave.org> |
|---|---|
| date | Wed, 16 Nov 2011 00:55:06 -0500 |
| parents | e90718520560 |
| children | |
| files | lrCostFunction.m predict.m |
| diffstat | 2 files changed, 3 insertions(+), 2 deletions(-) [+] |
line wrap: on
line diff
--- a/lrCostFunction.m +++ b/lrCostFunction.m @@ -11,7 +11,8 @@ ## h_theta(x) ht = sigmoid (X*theta); - J = -(y'*log (ht) + (1 - y)'*log (1 - ht))/m + lambda*sum (theta(2:end).^2)/(2*m); + J = -(y'*log (ht) + (1 - y)'*log (1 - ht))/m \ + + lambda*sumsq (theta(2:end))/(2*m); grad = (X'*(ht - y) + [0; lambda*theta(2:end,:)])/m ;
--- a/predict.m +++ b/predict.m @@ -6,7 +6,7 @@ m = rows (X); X = [ones(m, 1), X]; - ## See predictOneVsAll.m for an explanatin of this syntax if it's new + ## See predictOneVsAll.m for an explanation of this syntax if it's new ## to you. [~, p] = max ( [ones(m, 1), sigmoid(X*Theta1')]*Theta2', [], 2);