/* nag_tsa_multi_inp_model_estim (g13bec) Example Program.
*
* Copyright 2017 Numerical Algorithms Group.
*
* Mark 26.1, 2017.
*/
#include <nag.h>
#include <stdio.h>
#include <string.h>
#include <nag_string.h>
#include <nag_stdlib.h>
#include <nagg13.h>
#define XXY(I, J) xxy[(I) *tdxxy + J]
int main(void)
{
Integer exit_status = 0;
Integer i, inser, j, npara, nseries, nxxy, tdxxy;
Nag_ArimaOrder arimav;
Nag_G13_Opt options;
Nag_TransfOrder transfv;
double df, objf, *para = 0, rss, *sd = 0, *xxy = 0;
NagError fail;
INIT_FAIL(fail);
printf("nag_tsa_multi_inp_model_estim (g13bec) Example Program Results\n");
scanf(" %*[^\n]"); /* Skip heading in data file */
#define CM(I, J) options.cm[(J)+(I) *options.tdcm]
#define ZT(I, J) options.zt[(J)+(I) *options.tdzt]
/*
* Initialize the option structure.
*/
/* nag_tsa_options_init (g13bxc).
* Initialization function for option setting
*/
nag_tsa_options_init(&options);
scanf("%" NAG_IFMT "%" NAG_IFMT "%" NAG_IFMT "", &nxxy, &nseries,
&options.max_iter);
if (nxxy > 0 && nseries > 0) {
/*
* Set some specific option variables to the desired values.
*/
options.criteria = Nag_Marginal;
options.print_level = Nag_Soln_Iter_Full;
/*
* Allocate memory to the arrays in structure transfv containing
* the transfer function model orders of the input series.
*/
/* nag_tsa_transf_orders (g13byc), see above. */
nag_tsa_transf_orders(nseries, &transfv, &fail);
/*
* Read the orders vector of the ARIMA model for the output noise
* component into structure arimav.
*/
scanf("%" NAG_IFMT "%" NAG_IFMT "%" NAG_IFMT "%" NAG_IFMT "%" NAG_IFMT ""
"%" NAG_IFMT "%" NAG_IFMT "", &arimav.p, &arimav.d, &arimav.q,
&arimav.bigp, &arimav.bigd, &arimav.bigq, &arimav.s);
/*
* Read the transfer function model orders of the input series into
* structure transfv.
*/
inser = nseries - 1;
for (j = 0; j < inser; ++j)
scanf("%" NAG_IFMT "", &transfv.b[j]);
for (j = 0; j < inser; ++j)
scanf("%" NAG_IFMT "", &transfv.q[j]);
for (j = 0; j < inser; ++j)
scanf("%" NAG_IFMT "", &transfv.p[j]);
for (j = 0; j < inser; ++j)
scanf("%" NAG_IFMT "", &transfv.r[j]);
npara = 0;
for (i = 0; i < inser; ++i)
npara = npara + transfv.q[i] + transfv.p[i];
npara = npara + arimav.p + arimav.q + arimav.bigp + arimav.bigq + nseries;
if (npara >= 1) {
if (!(para = NAG_ALLOC(npara, double)) ||
!(sd = NAG_ALLOC(npara, double)) ||
!(xxy = NAG_ALLOC(nxxy * nseries, double)))
{
printf("Allocation failure\n");
exit_status = -1;
goto END;
}
tdxxy = nseries;
for (i = 0; i < npara; ++i)
scanf("%lf", ¶[i]);
for (i = 0; i < nxxy; ++i)
for (j = 0; j < nseries; ++j)
scanf("%lf", &XXY(i, j));
/* nag_tsa_multi_inp_model_estim (g13bec), see above. */
fflush(stdout);
nag_tsa_multi_inp_model_estim(&arimav, nseries, &transfv, para,
npara, nxxy, xxy, tdxxy, sd, &rss,
&objf, &df, &options, &fail);
if (fail.code != NE_NOERROR) {
printf("Error from nag_tsa_multi_inp_model_estim (g13bec)"
".\n%s\n", fail.message);
exit_status = 1;
goto END;
}
printf("\nThe correlation matrix is \n\n");
for (i = 0; i < npara; ++i)
for (j = 0; j < npara; ++j)
printf("%10.4f%c", CM(i, j), (j % 5 == 4) ? '\n' : ' ');
printf("\nThe residuals and the z and n values are\n\n");
printf(" i res[i] z(t) noise(t)\n\n");
for (i = 0; i < nxxy; ++i) {
if (i + 1 <= options.lenres) {
printf("%4" NAG_IFMT "%15.3f", i + 1, options.res[i]);
for (j = 0; j < nseries - 1; ++j)
printf("%15.3f ", ZT(i, j));
printf("%15.3f\n", options.noise[i]);
}
}
}
else {
printf("npara is out of range: npara = %-3" NAG_IFMT "\n", npara);
/* nag_tsa_free (g13xzc).
* Freeing function for use with g13 option setting
*/
nag_tsa_free(&options);
/* nag_tsa_trans_free (g13bzc), see above. */
nag_tsa_trans_free(&transfv);
exit_status = 1;
goto END;
}
}
else {
printf("One or both of nxxy and nseries are out of range:"
" nxxy = %-3" NAG_IFMT " while nseries = %-3" NAG_IFMT "\n",
nxxy, nseries);
exit_status = 1;
goto END;
}
/* nag_tsa_trans_free (g13bzc), see above. */
nag_tsa_trans_free(&transfv);
/* nag_tsa_free (g13xzc), see above. */
nag_tsa_free(&options);
END:
NAG_FREE(para);
NAG_FREE(sd);
NAG_FREE(xxy);
return exit_status;
}