/
svm.m
45 lines (34 loc) · 1.03 KB
/
svm.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
42
43
44
45
function [ b, b0, L ] = svm( X, Y, L_vec, K )
%SVM Compute linear support vector machine
%% Set default inputs and initialize variables
if (nargin == 2)
L_vec = logspace(-5,1,10);
K = 5;
elseif (nargin == 3)
K = 5;
end
[n, p] = size(X);
%% Generate K-Folds
ndK = floor(n/K);
fold_inds = randperm(n);
fold_inds = fold_inds(1:K*ndK);
fold_inds = reshape(fold_inds, ndK, K);
%% Perform K-Fold cross validation
cv_err = zeros(1, length(L_vec));
for ind = 1:K
train_inds = fold_inds(:, setdiff(1:K, ind));
train_inds = train_inds(:);
test_inds = fold_inds(:, ind);
pind_test = (Y(test_inds) >= 0);
nind_test = (Y(test_inds) <= 0);
for ind_j = 1:length(L_vec)
[b, b0] = solve_svm(X(train_inds,:), Y(train_inds), L_vec(ind_j));
vn = (X(test_inds,:)*b + b0 >= 0);
cv_err(ind_j) = cv_err(ind_j) + sum(1-vn(pind_test))/sum(pind_test) + sum(vn(nind_test))/sum(nind_test);
end
end
%% Compute the final SVM
[~, mind] = min(cv_err);
L = L_vec(mind);
[b, b0] = solve_svm(X, Y, L);
end