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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
|
/**
* @file gensvm_pred.c
* @author Gertjan van den Burg
* @date August 9, 2013
* @brief Main functions for predicting class labels..
*
* @details
* This file contains functions for predicting the class labels of instances
* and a function for calculating the predictive performance (hitrate) of
* a prediction given true class labels.
*
*/
#include <cblas.h>
#include "libGenSVM.h"
#include "gensvm.h"
#include "gensvm_kernel.h"
#include "gensvm_matrix.h"
#include "gensvm_pred.h"
#include "util.h" // testing
void gensvm_predict_labels(struct GenData *data_test,
struct GenData *data_train, struct GenModel *model,
long *predy)
{
if (model->kerneltype == K_LINEAR)
gensvm_predict_labels_linear(data_test, model, predy);
else
gensvm_predict_labels_kernel(data_test, data_train, model,
predy);
}
/**
* @brief Predict class labels of data given and output in predy
*
* @details
* The labels are predicted by mapping each instance in data to the
* simplex space using the matrix V in the given model. Next, for each
* instance the nearest simplex vertex is determined using an Euclidean
* norm. The nearest simplex vertex determines the predicted class label,
* which is recorded in predy.
*
* @param[in] data GenData to predict labels for
* @param[in] model GenModel with optimized V
* @param[out] predy pre-allocated vector to record predictions in
*/
void gensvm_predict_labels_linear(struct GenData *data,
struct GenModel *model, long *predy)
{
long i, j, k, label;
double norm, min_dist;
long n = data->n; // note that model->n is the size of the training sample.
long m = data->m;
long K = model->K; //data->K does not necessarily equal the original K.
double *S = Calloc(double, K-1);
double *ZV = Calloc(double, n*(K-1));
double *U = Calloc(double, K*(K-1));
// Get the simplex matrix
gensvm_simplex_gen(K, U);
// Generate the simplex-space vectors
cblas_dgemm(
CblasRowMajor,
CblasNoTrans,
CblasNoTrans,
n,
K-1,
m+1,
1.0,
data->Z,
m+1,
model->V,
K-1,
0.0,
ZV,
K-1);
// Calculate the distance to each of the vertices of the simplex.
// The closest vertex defines the class label.
for (i=0; i<n; i++) {
label = 0;
min_dist = 1000000000.0;
for (j=0; j<K; j++) {
for (k=0; k<K-1; k++) {
S[k] = matrix_get(ZV, K-1, i, k) -
matrix_get(U, K-1, j, k);
}
norm = cblas_dnrm2(K-1, S, 1);
if (norm < min_dist) {
label = j+1; // labels start counting from 1
min_dist = norm;
}
}
predy[i] = label;
}
free(ZV);
free(U);
free(S);
}
void gensvm_predict_labels_kernel(struct GenData *data_test,
struct GenData *data_train, struct GenModel *model,
long *predy)
{
long i, j, k, label;
double norm, min_dist;
long n_train = data_train->n;
long n_test = data_test->n;
long r = model->m;
long K = model->K;
double *K2 = NULL;
gensvm_make_crosskernel(model, data_train, data_test, &K2);
double *S = Calloc(double, K-1);
double *ZV = Calloc(double, n_test*(r+1));
double *KPS = Calloc(double, n_test*(r+1));
double *U = Calloc(double, K*(K-1));
gensvm_simplex_gen(K, U);
// were doing the computations explicitly since P is included in
// data_train->Z. Might want to look at this some more if it turns out
// to be slow.
double value, rowvalue;
for (i=0; i<n_test; i++) {
for (j=1; j<r+1; j++) {
value = 0.0;
for (k=0; k<n_train; k++) {
rowvalue = matrix_get(K2, n_train, i, k);
rowvalue *= matrix_get(data_train->Z, r+1, k,
j);
value += rowvalue;
}
value *= matrix_get(data_train->J, 1, j, 0);
matrix_set(KPS, r+1, i, j, value);
}
matrix_set(KPS, r+1, i, 0, 1.0);
}
cblas_dgemm(
CblasRowMajor,
CblasNoTrans,
CblasNoTrans,
n_test,
K-1,
r+1,
1.0,
KPS,
r+1,
model->V,
K-1,
0.0,
ZV,
K-1);
for (i=0; i<n_test; i++) {
label = 0;
min_dist = 1e10;
for (j=0; j<K; j++) {
for (k=0; k<K-1; k++) {
S[k] = matrix_get(ZV, K-1, i, k) -
matrix_get(U, K-1, j, k);
}
norm = cblas_dnrm2(K, S, 1);
if (norm < min_dist) {
label = j+1;
min_dist = norm;
}
}
predy[i] = label;
}
free(ZV);
free(U);
free(S);
free(KPS);
free(K2);
}
/**
* @brief Calculate the predictive performance (percentage correct)
*
* @details
* The predictive performance is calculated by simply counting the number
* of correctly classified samples and dividing by the total number of
* samples, multiplying by 100.
*
* @param[in] data the GenData dataset with known labels
* @param[in] predy the predicted class labels
*
* @returns percentage correctly classified.
*/
double gensvm_prediction_perf(struct GenData *data, long *predy)
{
long i, correct = 0;
double performance;
for (i=0; i<data->n; i++)
if (data->y[i] == predy[i])
correct++;
performance = ((double) correct)/((double) data->n)* 100.0;
return performance;
}
|
