Mercurial > hg > tvii
changeset 45:4d173452377e
notes on backpropagation
author | Jeff Hammel <k0scist@gmail.com> |
---|---|
date | Mon, 04 Sep 2017 15:19:29 -0700 |
parents | 857a606783e1 |
children | cb1a02a413bc |
files | tvii/logistic_regression.py |
diffstat | 1 files changed, 5 insertions(+), 0 deletions(-) [+] |
line wrap: on
line diff
--- a/tvii/logistic_regression.py Mon Sep 04 15:06:38 2017 -0700 +++ b/tvii/logistic_regression.py Mon Sep 04 15:19:29 2017 -0700 @@ -15,6 +15,11 @@ from .sigmoid import sigmoid +def loss(a, y): + # UNTESTED! + # derivative = -(y/a) + (1-y)/(1-a) + return -y*np.log(a) - (1-y)*np.log(1-a) + def propagate(w, b, X, Y): """ Implement the cost function and its gradient for the propagation: