void vec_scaleTo(Vector2* vec, float newLength, Vector2* vDest)
{
float oldLength = vec_length2(vec);
if (oldLength != 0)
{
vec_mult_scalar(vec, newLength/oldLength, vDest);
}
}
/*
* Calculate desired racket velocity given ball incoming and
* outgoing velocities
* Assuming a mirror law
* Assumes no desired spin, i.e. racket velocity along the racket will be set to zero
*
* Output is the last parameter: racketVel
*
* TODO: is the CRR value accurate?
*
*/
void calc_racket_vel(Vector velBallIn, Vector velBallOut, Vector normalRacket, Vector velRacket) {
double velBallInAlongNormal;
double velBallOutAlongNormal;
double eps = CRR;
velBallInAlongNormal = vec_mult_inner(velBallIn, normalRacket);
velBallOutAlongNormal = vec_mult_inner(velBallOut, normalRacket);
velBallInAlongNormal = eps * velBallInAlongNormal;
velBallOutAlongNormal = (velBallOutAlongNormal + velBallInAlongNormal) / (1+eps);
vec_mult_scalar(normalRacket, velBallOutAlongNormal, velRacket);
}
/*!*****************************************************************************
*******************************************************************************
\note parm_opt
\date 10/20/91
\remarks
this is the major optimzation program
*******************************************************************************
Function Parameters: [in]=input,[out]=output
\param[in] tol : error tolernance to be achieved
\param[in] n_parm : number of parameters to be optimzed
\param[in] n_con : number of contraints to be taken into account
\param[in] f_dLda : function which calculates the derivative of the
optimziation criterion with respect to the parameters;
must return vector
\param[in] f_dMda : function which calculates the derivate of the
constraints with respect to parameters
must return matrix
\param[in] f_M : constraints function, must always be formulted to
return 0 for properly fulfilled constraints
\param[in] f_L : function to calculate simple cost (i.e., constraint
cost NOT included), the constraint costs are added
by this program automatically, the function returns
a double scalar
\param[in] f_dLdada : second derivative of L with respect to the parameters,
must be a matrix of dim n_parm x n_parm
\param[in] f_dMdada : second derivative of M with respect to parameters,
must be a matrix n_con*n_parm x n_parm
\param[in] use_newton: TRUE or FALSE to indicate that second derivatives
are given and can be used for Newton algorithm
\param[in,out] a : initial setting of parameters and also return of
optimal value (must be a vector, even if scalar)
\param[out] final_cost: the final cost
\param[out] err : the sqrt of the squared error of all constraints
NOTE: - program returns TRUE if everything correct, otherwise FALSE
- always minimizes the cost!!!
- algorithms come from Dyer McReynolds
NOTE: besides the possiblity of a bug, the Newton method seems to
sacrifice the validity of the constraint a little up to
quite a bit and should be used prudently
******************************************************************************/
int
parm_opt(double *a,int n_parm, int n_con, double *tol, void (*f_dLda)(),
void (*f_dMda)(), void (*f_M)(), double (*f_L)(), void (*f_dMdada)(),
void (*f_dLdada)(), int use_newton, double *final_cost, double *err)
{
register int i,j,n;
double cost= 999.e30;
double last_cost = 0.0;
double *mult=NULL, *new_mult=NULL; /* this is the vector of Lagrange mulitplier */
double **dMda=NULL, **dMda_t=NULL;
double *dLda;
double *K=NULL; /* the error in the constraints */
double eps = 0.025; /* the learning rate */
double **aux_mat=NULL; /* needed for inversion of matrix */
double *aux_vec=NULL;
double *new_a;
double **dMdada=NULL;
double **dLdada=NULL;
double **A=NULL; /* big matrix, a combination of several other matrices */
double *B=NULL; /* a big vector */
double **A_inv=NULL;
int rc=TRUE;
long count = 0;
int last_sign = 1;
int pending1 = FALSE, pending2 = FALSE;
int firsttime = TRUE;
int newton_active = FALSE;
dLda = my_vector(1,n_parm);
new_a = my_vector(1,n_parm);
if (n_con > 0) {
mult = my_vector(1,n_con);
dMda = my_matrix(1,n_con,1,n_parm);
dMda_t = my_matrix(1,n_parm,1,n_con);
K = my_vector(1,n_con);
aux_mat = my_matrix(1,n_con,1,n_con);
aux_vec = my_vector(1,n_con);
}
if (use_newton) {
dLdada = my_matrix(1,n_parm,1,n_parm);
A = my_matrix(1,n_parm+n_con,1,n_parm+n_con);
A_inv = my_matrix(1,n_parm+n_con,1,n_parm+n_con);
B = my_vector(1,n_parm+n_con);
if (n_con > 0) {
dMdada = my_matrix(1,n_con*n_parm,1,n_parm);
new_mult = my_vector(1,n_con);
}
for (i=1+n_parm; i<=n_con+n_parm; ++i) {
for (j=1+n_parm; j<=n_con+n_parm; ++j) {
A[i][j] = 0.0;
}
}
}
while (fabs(cost-last_cost) > *tol) {
++count;
pending1 = FALSE;
pending2 = FALSE;
AGAIN:
/* calculate the current Lagrange multipliers */
if (n_con > 0) {
(*f_M)(a,K); /* takes the parameters, returns residuals */
(*f_dMda)(a,dMda); /* takes the parameters, returns the Jacobian */
}
(*f_dLda)(a,dLda); /* takes the parameters, returns the gradient */
if (n_con > 0) {
mat_trans(dMda,dMda_t);
}
if (newton_active) {
if (n_con > 0) {
(*f_dMdada)(a,dMdada);
}
(*f_dLdada)(a,dLdada);
}
/* the first step is always a gradient step */
if (newton_active) {
if (firsttime) {
firsttime = FALSE;
eps = 0.1;
}
/* build the A matrix */
for (i=1; i<=n_parm; ++i) {
for (j=1; j<=n_parm; ++j) {
A[i][j] = dLdada[i][j];
for (n=1; n<=n_con; ++n) {
A[i][j] += mult[n]*dMdada[n+(i-1)*n_con][j];
}
}
}
for (i=1+n_parm; i<=n_con+n_parm; ++i) {
for (j=1; j<=n_parm; ++j) {
A[j][i] = A[i][j] = dMda[i-n_parm][j];
}
}
/* build the B vector */
if (n_con > 0) {
mat_vec_mult(dMda_t,mult,B);
}
for (i=1; i<=n_con; ++i) {
B[i+n_parm] = K[i];
}
/* invert the A matrix */
if (!my_inv_ludcmp(A, n_con+n_parm, A_inv)) {
rc = FALSE;
break;
}
mat_vec_mult(A_inv,B,B);
vec_mult_scalar(B,eps,B);
for (i=1; i<=n_parm; ++i) {
new_a[i] = a[i] + B[i];
}
for (i=1; i<=n_con; ++i) {
new_mult[i] = mult[i] + B[n_parm+i];
}
} else {
if (n_con > 0) {
/* the mulitpliers are updated according:
mult = (dMda dMda_t)^(-1) (K/esp - dMda dLda_t) */
mat_mult(dMda,dMda_t,aux_mat);
if (!my_inv_ludcmp(aux_mat, n_con, aux_mat)) {
rc = FALSE;
break;
}
mat_vec_mult(dMda,dLda,aux_vec);
vec_mult_scalar(K,1./eps,K);
vec_sub(K,aux_vec,aux_vec);
mat_vec_mult(aux_mat,aux_vec,mult);
}
/* the update step looks the following:
a_new = a - eps * (dLda + mult_t * dMda)_t */
if (n_con > 0) {
vec_mat_mult(mult,dMda,new_a);
vec_add(dLda,new_a,new_a);
} else {
vec_equal(dLda,new_a);
}
vec_mult_scalar(new_a,eps,new_a);
vec_sub(a,new_a,new_a);
}
if (count == 1 && !pending1) {
last_cost = (*f_L)(a);
if (n_con > 0) {
(*f_M)(a,K);
last_cost += vec_mult_inner(K,mult);
}
} else {
last_cost = cost;
}
/* calculate the updated cost */
cost = (*f_L)(new_a);
/*printf(" %f\n",cost);*/
if (n_con > 0) {
(*f_M)(new_a,K);
if (newton_active) {
cost += vec_mult_inner(K,new_mult);
} else {
cost += vec_mult_inner(K,mult);
}
}
/* printf("last=%f new=%f\n",last_cost,cost); */
/* check out whether we reduced the cost */
if (cost > last_cost && fabs(cost-last_cost) > *tol) {
/* reduce the gradient climbing rate: sometimes a reduction of eps
causes an increase in cost, thus leave an option to increase
eps */
cost = last_cost; /* reset last_cost */
if (pending1 && pending2) {
/* this means that either increase nor decrease
of eps helps, ==> leave the program */
rc = TRUE;
break;
} else if (pending1) {
eps *= 4.0; /* the last cutting by half did not help, thus
multiply by 2 to get to previous value, and
one more time by 2 to get new value */
pending2 = TRUE;
} else {
eps /= 2.0;
pending1 = TRUE;
}
goto AGAIN;
} else {
vec_equal(new_a,a);
if (newton_active && n_con > 0) {
vec_equal(new_mult,mult);
}
if (use_newton && fabs(cost-last_cost) < NEWTON_THRESHOLD)
newton_active = TRUE;
}
}
my_free_vector(dLda,1,n_parm);
my_free_vector(new_a,1,n_parm);
if (n_con > 0) {
my_free_vector(mult,1,n_con);
my_free_matrix(dMda,1,n_con,1,n_parm);
my_free_matrix(dMda_t,1,n_parm,1,n_con);
my_free_vector(K,1,n_con);
my_free_matrix(aux_mat,1,n_con,1,n_con);
my_free_vector(aux_vec,1,n_con);
}
if (use_newton) {
my_free_matrix(dLdada,1,n_parm,1,n_parm);
my_free_matrix(A,1,n_parm+n_con,1,n_parm+n_con);
my_free_matrix(A_inv,1,n_parm+n_con,1,n_parm+n_con);
my_free_vector(B,1,n_parm+n_con);
if (n_con > 0) {
my_free_matrix(dMdada,1,n_con*n_parm,1,n_parm);
my_free_vector(new_mult,1,n_con);
}
}
*final_cost = cost;
*tol = fabs(cost-last_cost);
if (n_con > 0) {
*err = sqrt(vec_mult_inner(K,K));
} else {
*err = 0.0;
}
/*
printf("count=%ld rc=%d\n",count,rc);
*/
return rc;
}
/*
* Calculate desired racket normal using the mirror law
*/
void calc_racket_normal(Vector bin, Vector bout, Vector normal) {
vec_sub(bout, bin, normal);
// normalize
vec_mult_scalar(normal, 1./sqrt(vec_mult_inner(normal,normal)), normal);
}
void
read_script(void)
{
int i,j,k,m,rc;
char fname[100];
double dummy;
int idummy;
FILE *in,*out;
int n_train_data_columns;
int n_test_data_columns;
Vector row;
char identification_string[100];
char string[100];
int num;
char vnames[50][100];
int ans = 0;
double o_noise, c_noise;
int old_indx_flag = FALSE;
Matrix D_train=NULL;
char **vnames_train=NULL;
char **units_train=NULL;
double freq_train;
int n_cols_train;
int n_rows_train;
Matrix D_test=NULL;
char **vnames_test=NULL;
char **units_test=NULL;
double freq_test;
int n_cols_test;
int n_rows_test;
/* I need the filename of the script file: first check whether the
user provided it in the argv_global variables */
if (argc_global > 0) {
in = fopen(argv_global[1],"r");
if (in != NULL) {
fclose(in);
strcpy(fname,argv_global[1]);
} else {
if (!getFile(fname)) exit(-1);
}
} else {
if (!getFile(fname)) exit(-1);
}
/* this allows to generate the LWPR */
if (argc_global > 1) {
sscanf(argv_global[2],"%d",&new_model);
} else {
get_int("Generate new LWPR = 1; Read from file = 0",ans,&ans);
if (ans) new_model = TRUE;
}
if (!readLWPRScript(fname,new_model,LWPR1)) {
fprintf(stderr,"Error when reading script file >%s<\n",fname);
exit(-1);
}
/* now read additional variables from the script */
in = fopen_strip(fname);
/* check for included files */
if (find_keyword(in,"include")) {
char fname_include[100];
FILE *fp_include;
rc=fscanf(in,"%s",fname_include);
fp_include = fopen_strip(fname_include);
fseek(fp_include, 0, SEEK_END);
rewind(in);
while ((rc=fgetc(in)) != EOF)
fputc(rc,fp_include);
fclose(in);
in = fp_include;
rewind(in);
}
/* All the names in file will be parsed. Here I define the names of the
variables. Note that the variables defining the dimensionality of
the LWPR must come first since we need them to allocate other
variables */
i=0;
/* this block has all variables needed to create a LWPR */
sprintf(vnames[++i],"n_in_w");
sprintf(vnames[++i],"n_in_reg");
sprintf(vnames[++i],"n_out");
sprintf(vnames[++i],"lwpr_name");
sprintf(vnames[++i],"n_in_reg_2nd");
sprintf(vnames[++i],"n_out_2nd");
/* this block specifies all variables needed to get the data
for training and testing, as well as some other parameters */
sprintf(vnames[++i],"sampling_method");
sprintf(vnames[++i],"index_function");
sprintf(vnames[++i],"max_iterations");
sprintf(vnames[++i],"eval_time");
sprintf(vnames[++i],"cutoff");
sprintf(vnames[++i],"blending");
sprintf(vnames[++i],"file_name_train_data");
sprintf(vnames[++i],"file_name_test_data");
sprintf(vnames[++i],"name_train_in_w_columns");
sprintf(vnames[++i],"name_train_in_reg_columns");
sprintf(vnames[++i],"name_train_out_columns");
sprintf(vnames[++i],"name_test_in_w_columns");
sprintf(vnames[++i],"name_test_in_reg_columns");
sprintf(vnames[++i],"name_test_out_columns");
sprintf(vnames[++i],"name_train_in_reg_2nd_columns");
sprintf(vnames[++i],"name_train_out_2nd_columns");
sprintf(vnames[++i],"name_test_in_reg_2nd_columns");
sprintf(vnames[++i],"name_test_out_2nd_columns");
out = fopen("lwpr_test_varnames","w");
for (j=1; j<=i; ++j)
fprintf(out,"%s\n",vnames[j]);
fclose(out);
remove_temp_file();
/* parse keywords */
i = 0;
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
rc=fscanf(in,"%d",&n_in_w);
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
rc=fscanf(in,"%d",&n_in_reg);
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
rc=fscanf(in,"%d",&n_out);
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
rc=fscanf(in,"%s",lwpr_name);
if (!find_keyword(in,vnames[++i])) {
if (lwprs[LWPR1].use_reg_2nd) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
} else {
rc=fscanf(in,"%d",&n_in_reg_2nd);
}
if (!find_keyword(in,vnames[++i])) {
if (lwprs[LWPR1].use_reg_2nd) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
} else {
rc=fscanf(in,"%d",&n_out_2nd);
}
/* at last the parameters need to steer the training and testing of
the LWPR */
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
rc=fscanf(in,"%d",&sampling_method);
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
rc=fscanf(in,"%d",&index_function);
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
rc=fscanf(in,"%ld",&max_iterations);
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
rc=fscanf(in,"%ld",&eval_time);
if (find_keyword(in,vnames[++i])) {
rc=fscanf(in,"%lf",&cutoff);
}
if (find_keyword(in,vnames[++i])) {
rc=fscanf(in,"%d",&blending);
}
if (!find_keyword(in,vnames[++i]) && argc_global <= 2) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
if (argc_global > 2) {
strcpy(fname_train_data,argv_global[3]);
} else {
rc=fscanf(in,"%s",fname_train_data);
}
if (!find_keyword(in,vnames[++i]) && argc_global <= 3) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
if (argc_global > 3) {
strcpy(fname_test_data,argv_global[4]);
} else {
rc=fscanf(in,"%s",fname_test_data);
}
// at this point the data files can be read -- they are expected to be in
// MRDPLOT format
printf("Reading training data from >%s<...",fname_train_data);
if (!mrdplot_convert(fname_train_data, &D_train, &vnames_train, &units_train,
&freq_train, &n_cols_train, &n_rows_train)) {
printf("Problems reading MRDPLOT file >%s<\n",fname_train_data);
exit(-999);
}
printf("done\n");
printf("%d rows with %d columns read\n",n_rows_train,n_cols_train);
printf("Reading test data from >%s<...",fname_test_data);
if (!mrdplot_convert(fname_test_data, &D_test, &vnames_test, &units_test,
&freq_test, &n_cols_test, &n_rows_test)) {
printf("Problems reading MRDPLOT file >%s<\n",fname_test_data);
exit(-999);
}
printf("done\n");
printf("%d rows with %d columns read\n",n_rows_test,n_cols_test);
// allocate memory for all arrays
Xw_train = my_matrix(1,n_rows_train,1,n_in_w);
Xreg_train = my_matrix(1,n_rows_train,1,n_in_reg);
Xreg_train_2nd = my_matrix(1,n_rows_train,1,n_in_reg_2nd);
Xw_test = my_matrix(1,n_rows_test,1,n_in_w);
Xreg_test = my_matrix(1,n_rows_test,1,n_in_reg);
Xreg_test_2nd = my_matrix(1,n_rows_test,1,n_in_reg_2nd);
Y_train = my_matrix(1,n_rows_train,1,n_out);
Y_train_2nd = my_matrix(1,n_rows_train,1,n_out_2nd);
Y_test = my_matrix(1,n_rows_test,1,n_out);
Y_test_2nd = my_matrix(1,n_rows_test,1,n_out_2nd);
x_w = my_vector(1,n_in_w);
x_reg = my_vector(1,n_in_reg);
x_reg_2nd = my_vector(1,n_in_reg_2nd);
y = my_vector(1,n_out);
y_2nd = my_vector(1,n_out_2nd);
conf = my_vector(1,n_out);
conf_2nd = my_vector(1,n_out_2nd);
var_y = my_vector(1,n_out);
var_y_2nd = my_vector(1,n_out_2nd);
mean_y = my_vector(1,n_out);
mean_y_2nd = my_vector(1,n_out_2nd);
// sort the test and training data into the appropriate arrays
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_in_w; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_train,n_cols_train))) {
printf("Couldn't find column >%s< in training data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_train; ++m)
Xw_train[m][j] = D_train[m][k];
}
}
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_in_reg; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_train,n_cols_train))) {
printf("Couldn't find column >%s< in training data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_train; ++m)
Xreg_train[m][j] = D_train[m][k];
}
}
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_out; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_train,n_cols_train))) {
printf("Couldn't find column >%s< in training data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_train; ++m)
Y_train[m][j] = D_train[m][k];
}
}
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_in_w; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_test,n_cols_test))) {
printf("Couldn't find column >%s< in test data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_test; ++m)
Xw_test[m][j] = D_test[m][k];
}
}
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_in_reg; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_test,n_cols_test))) {
printf("Couldn't find column >%s< in test data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_test; ++m)
Xreg_test[m][j] = D_test[m][k];
}
}
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_out; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_test,n_cols_test))) {
printf("Couldn't find column >%s< in test data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_test; ++m)
Y_test[m][j] = D_test[m][k];
}
}
if (lwprs[LWPR1].use_reg_2nd) {
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_in_reg_2nd; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_train,n_cols_train))) {
printf("Couldn't find column >%s< in training data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_train; ++m)
Xreg_train_2nd[m][j] = D_train[m][k];
}
}
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_out_2nd; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_train,n_cols_train))) {
printf("Couldn't find column >%s< in training data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_train; ++m)
Y_train_2nd[m][j] = D_train[m][k];
}
}
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_in_reg_2nd; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_test,n_cols_test))) {
printf("Couldn't find column >%s< in test data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_test; ++m)
Xreg_test_2nd[m][j] = D_test[m][k];
}
}
if (!find_keyword(in,vnames[++i])) {
printf("Could not find variable >%s<\n",vnames[i]);
exit(-i);
}
for (j=1; j<=n_out_2nd; ++j) {
rc=fscanf(in,"%s",string);
if (!(k=findIndex(string,vnames_test,n_cols_test))) {
printf("Couldn't find column >%s< in test data\n",string);
exit(-i);
} else {
for (m=1; m<=n_rows_test; ++m)
Y_test_2nd[m][j] = D_test[m][k];
}
}
}
fclose(in);
n_train_data = n_rows_train;
n_test_data = n_rows_test;
if (NORMALIZE_BY_TRAIN) {
for (i=1; i<=n_train_data; ++i) {
vec_add_size(mean_y,Y_train[i],n_out,mean_y);
}
vec_mult_scalar(mean_y,1./(double)n_train_data,mean_y);
for (i=1; i<=n_train_data; ++i) {
for (j=1; j<=n_out; ++j) {
var_y[j] += sqr(Y_train[i][j]-mean_y[j]);
}
}
vec_mult_scalar(var_y,1./(double)n_train_data,var_y);
if (lwprs[LWPR1].use_reg_2nd) {
for (i=1; i<=n_train_data; ++i) {
vec_add_size(mean_y_2nd,Y_train_2nd[i],n_out,mean_y_2nd);
}
vec_mult_scalar(mean_y_2nd,1./(double)n_train_data,mean_y_2nd);
for (i=1; i<=n_train_data; ++i) {
for (j=1; j<=n_out_2nd; ++j) {
var_y_2nd[j] += sqr(Y_train_2nd[i][j]-mean_y_2nd[j]);
}
}
vec_mult_scalar(var_y_2nd,1./(double)n_train_data,var_y_2nd);
}
} else {
for (i=1; i<=n_test_data; ++i) {
vec_add_size(mean_y,Y_test[i],n_out,mean_y);
}
vec_mult_scalar(mean_y,1./(double)n_test_data,mean_y);
for (i=1; i<=n_test_data; ++i) {
for (j=1; j<=n_out; ++j) {
var_y[j] += sqr(Y_test[i][j]-mean_y[j]);
}
}
vec_mult_scalar(var_y,1./(double)n_test_data,var_y);
if (lwprs[LWPR1].use_reg_2nd) {
for (i=1; i<=n_test_data; ++i) {
vec_add_size(mean_y_2nd,Y_test_2nd[i],n_out,mean_y_2nd);
}
vec_mult_scalar(mean_y_2nd,1./(double)n_test_data,mean_y_2nd);
for (i=1; i<=n_train_data; ++i) {
for (j=1; j<=n_out_2nd; ++j) {
var_y_2nd[j] += sqr(Y_test_2nd[i][j]-mean_y_2nd[j]);
}
}
vec_mult_scalar(var_y_2nd,1./(double)n_train_data,var_y_2nd);
}
}
}
