-
Notifications
You must be signed in to change notification settings - Fork 61
/
smooth_logLLogistic.m
41 lines (35 loc) · 1.34 KB
/
smooth_logLLogistic.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
function op = smooth_logLLogistic(y)
% SMOOTH_LOGLLOGISTIC Log-likelihood function of a logistic: sum_i( y_i mu_i - log( 1+exp(mu_i) ) )
% OP = SMOOTH_LOGLLOGISTIC( Y )
% returns a function that computes the log-likelihood function
% in a standard logistic regression model with independent entries. There
% are two classes y_i = 0 and y_i = 1 with
%
% prob(y_i = 1) = exp(mu_i)/(1 + exp(mu_i)
%
% so that the log-likelihood is given by
%
% log-likelihood(mu) = sum_i ( y_i mu_i - log(1+ exp(mu_i)) )
%
% where mu is the parameter of the distribution (this is unknown,
% so it is the variable), and Y is a vector of observations.
error(nargchk(1,1,nargin));
op = tfocs_smooth( @smooth_logLlogistic_impl);
function [ v, g ] = smooth_logLlogistic_impl( mu )
if length(mu) == 1,
mu = mu * ones(size(y));
elseif size(mu) ~= size(y),
error('Parameters and data must be of the same size'),
end
aux = 1 + exp(-abs(mu));
v = tfocs_dot(y-1,mu.*(mu > 0)) ...
+ tfocs_dot(y,mu.*(mu < 0)) ...
- tfocs_dot(ones(size(y)), log(aux));
if nargout > 1,
g = y - ((mu > 0) + (mu <= 0).*exp(mu))./aux;
end
end
end
% TFOCS v1.3 by Stephen Becker, Emmanuel Candes, and Michael Grant.
% Copyright 2013 California Institute of Technology and CVX Research.
% See the file LICENSE for full license information.