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
|
#include "ranksvmtn.h"
using namespace std;
using namespace Eigen;
const int maxiter = 10;
const double prec=1e-3;
int cg_solve(const MatrixXd &A, const VectorXd &b, const VectorXd &x)
{
double alpha,beta,r_1,r_2;
VectorXd p = x;
VectorXd q;
VectorXd res;
while (1)
{
beta = r_1/r_2;
p = res + beta*p;
q = A*p;
alpha = r_1/p.dot(q);
// Non preconditioned version
alpha = p.dot(p)/(p.dot(q));
res=res-alpha*q;
break;
}
return 0;
}
// Calculate objfunc gradient & support vectors
int objfunc_linear(const VectorXd &w,const double C,const VectorXd &pred,const VectorXd &grad, double &obj,MatrixXd &sv)
{
pred = pred.cwiseMax(Matrix::Zero(pred.rows(),pred.cols()));
obj = (pred.cwiseProduct(pred)*(C/2)) + w.transpose()*w/2;
grad = w - (((pred*C).transpose()*A)*w).transpose();
for (int i=0;i<pred.cols();++i)
if (pred(i)>0)
sv(i,i)=1;
else
sv(i,i)=0;
return 0;
}
// line search
int line_search(const VectorXd &w,const double C,const VectorXd &step,VectorXd &pred,double &t)
{
return 0;
}
int RSVMTN::train(DataSet &D, Labels &label){
int iter = 0;
MatrixXd A;
int n=D.rows();
LOG(INFO) << "training with feature size:" << fsize << " Data size:" << n;
MatrixXd sv=MatrixXd::Identity(n, n);
VectorXd grad(fsize);
VectorXd step(fsize);
VectorXd pred(n);
double obj,t;
pred=VectorXd::Ones(n) - (A*(D*model.weight));
while (true)
{
iter+=1;
if (iter> maxiter)
{
LOG(INFO)<< "Maxiter :"<<maxiter<<" reached";
break;
}
// Generate support vector matrix sv & gradient
objfunc_linear(D,1,pred,grad,obj,sv);
model.weight=model.weight+step*t;
// When dec is small enough
if (-step.dot(grad) < prec * obj)
break;
}
return 0;
};
int RSVMTN::predict(DataSet &D, Labels &res){
res = model.weight * D;
for (int i=0;i<res.cols();++i)
res[i] = (res[i] + model.beta);
return 0;
};
|
