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
|
/*
* train_logReg_param.hh
*
* License: Artistic License, see file LICENSE.TXT or
* https://opensource.org/licenses/artistic-license-1.0
*/
#ifndef _TRAINLOGREG_HH
#define _TRAINLOGREG_HH
// project includes
#include "types.hh"
#include "properties.hh"
// standard C/C++ includes
#include <iostream>
#include <vector>
#include <unordered_map>
#include <gsl/gsl_multimin.h>
using namespace std;
/**
* @brief implementation of logistic regression for training OE and EC
* feature parameters using GSL
*
* @author Lizzy Gerischer
*/
class train_data{
public:
train_data(unordered_map<string, pair<int, vector<double> > > *samples, unordered_map<string,int> *ref_class, int numSp);
vector<pair<int, vector<double> > > *exon_samples;
vector<pair<int, vector<double> > > *intron_samples;
int num_features;
double prob_true_false_exons;
double prob_true_false_introns;
vector<int> exon_feature_idx;
vector<int> intron_feature_idx;
void set_mean_std(vector<pair<int, vector<double> > > *samples, vector<int> feature_idx);
};
double activation_f(const gsl_vector *theta, vector<double> *f);
double get_feature(vector<double> *f, int i);
double cross_entropy_error_f(const gsl_vector *theta, void *features);
void cross_entropy_error_df(const gsl_vector *theta, void *features, gsl_vector *df);
void cross_entropy_error_fdf(const gsl_vector *params, void *features, double *f, gsl_vector *df);
double rob_cross_entropy_error_f(const gsl_vector *theta, void *features);
void rob_cross_entropy_error_df(const gsl_vector *theta, void *features, gsl_vector *df);
void rob_cross_entropy_error_fdf(const gsl_vector *params, void *features, double *f, gsl_vector *df);
void optimize_parameters(train_data *data);
void reference_from_file(unordered_map<string,int> *ref_class);
void trainFeature_from_file();
void train_OEscore_params(int numSpecies);
void sgd(vector<pair<int, vector<double> > > *samples, gsl_vector *theta, int n);
void gsl_minimizer(vector<pair<int, vector<double> > > *samples, gsl_vector *theta, int n);
#endif
|
