NAG Library Manual, Mark 28.7
Interfaces:  FL   CL   CPP   AD 

NAG CL Interface Introduction
Example description
/* nag_correg_ridge (g02kbc) Example Program.
 *
 * Copyright 2022 Numerical Algorithms Group.
 *
 * Mark 28.7, 2022.
 */
/* Pre-processor includes */
#include <math.h>
#include <nag.h>
#include <stdio.h>

int main(void) {
  /*Integer scalar and array declarations */
  Integer exit_status = 0;
  Integer i, ip, ip1, j, lh, lpec, m, n, pl;
  Integer pdb, pdpe, pdvf, pdx;
  Integer *isx = 0;
  /*Double scalar and array declarations */
  double *b = 0, *h = 0, *nep = 0, *pe = 0, *vf = 0, *x = 0, *y = 0;
  /*Character scalar and array declarations */
  char spec[40], swantb[40];
  /*NAG Types */
  Nag_OrderType order;
  Nag_ParaOption wantb;
  Nag_VIFOption wantvf;
  Nag_PredictError *pec = 0;
  NagError fail;

  INIT_FAIL(fail);

  printf("%s\n", "nag_correg_ridge (g02kbc) Example Program Results");
  /* Skip heading in data file */
  scanf("%*[^\n] ");
  /* Read in the problem size information */
  scanf("%" NAG_IFMT "%" NAG_IFMT "%" NAG_IFMT "%" NAG_IFMT "%39s%*[^\n] ", &n,
        &m, &lh, &lpec, swantb);
  wantb = (Nag_ParaOption)nag_enum_name_to_value(swantb);
  if (!(isx = NAG_ALLOC(m, Integer))) {
    printf("Allocation failure\n");
    exit_status = -1;
    goto END;
  }
  /* Read in the ISX flags */
  for (i = 0; i < m; i++)
    scanf("%" NAG_IFMT " ", &isx[i]);
  scanf("%*[^\n] ");
  /* Total number of variables */
  ip = 0;
  for (j = 0; j < m; j++) {
    if (isx[j] == 1)
      ip = ip + 1;
  }
#ifdef NAG_COLUMN_MAJOR
  pdb = ip + 1;
#define B(I, J) b[(J - 1) * pdb + I - 1]
  pdpe = lpec;
#define PE(I, J) pe[(J - 1) * pdpe + I - 1]
  pdvf = ip;
#define VF(I, J) vf[(J - 1) * pdvf + I - 1]
  pdx = n;
#define X(I, J) x[(J - 1) * pdx + I - 1]
  order = Nag_ColMajor;
#else
  pdb = lh;
#define B(I, J) b[(I - 1) * pdb + J - 1]
  pdpe = lh;
#define PE(I, J) pe[(I - 1) * pdpe + J - 1]
  pdvf = lh;
#define VF(I, J) vf[(I - 1) * pdvf + J - 1]
  pdx = m;
#define X(I, J) x[(I - 1) * pdx + J - 1]
  order = Nag_RowMajor;
#endif
  if (!(b = NAG_ALLOC(pdb * (order == Nag_RowMajor ? (ip + 1) : lh), double)) ||
      !(h = NAG_ALLOC(lh, double)) || !(nep = NAG_ALLOC(lh, double)) ||
      !(pe = NAG_ALLOC(pdpe * (order == Nag_RowMajor ? lpec : lh), double)) ||
      !(vf = NAG_ALLOC(pdvf * (order == Nag_RowMajor ? ip : lh), double)) ||
      !(x = NAG_ALLOC(pdx * (order == Nag_RowMajor ? n : m), double)) ||
      !(y = NAG_ALLOC(n, double)) ||
      !(pec = NAG_ALLOC(lpec, Nag_PredictError))) {
    printf("Allocation failure\n");
    exit_status = -1;
    goto END;
  }
  /* Read in the data */
  if (lpec > 0) {
    for (i = 0; i < lpec; i++) {
      scanf("%39s ", spec);
      pec[i] = (Nag_PredictError)nag_enum_name_to_value(spec);
    }
    scanf("%*[^\n] ");
  }
  for (i = 1; i <= n; i++) {
    for (j = 1; j <= m; j++)
      scanf("%lf ", &X(i, j));
    scanf("%lf ", &y[i - 1]);
  }
  scanf("%*[^\n] ");
  /*  Read in the ridge coefficients */
  for (i = 0; i < lh; i++)
    scanf("%lf ", &h[i]);
  scanf("%*[^\n] ");
  /* Output the variance inflation factors and parameter estimates */
  wantvf = Nag_WantVIF;
  /* Run the analysis */
  /*
   * nag_correg_ridge (g02kbc)
   * Ridge regression
   */
  nag_correg_ridge(order, n, m, x, pdx, isx, ip, y, lh, h, nep, wantb, b, pdb,
                   wantvf, vf, pdvf, lpec, pec, pe, pdpe, &fail);
  if (fail.code != NE_NOERROR) {
    printf("Error from nag_correg_ridge (g02kbc).\n%s\n", fail.message);
    exit_status = 1;
    goto END;
  }
  /* Output results */
  ip1 = ip - 1;
  /* Summaries */
  printf("%s%10" NAG_IFMT "\n", "Number of parameters used = ", ip + 1);
  printf("%s\n", "Effective number of parameters (NEP):");
  printf("%s\n", "   Ridge   ");
  printf("%s%s\n", "   Coeff.  ", "NEP");
  for (i = 1; i <= lh; i++)
    printf(" %10.4f %10.4f\n", h[i - 1], nep[i - 1]);
  /* Parameter estimates */
  if (wantb != Nag_NoPara) {
    printf("\n");
    if (wantb == Nag_OrigPara) {
      printf("%s\n", "Parameter Estimates (Original scalings)");
    } else {
      printf("%s\n", "Parameter Estimates (Standarised)");
    }
    pl = MIN(ip, 4);
    printf("%s\n", "  Ridge  ");
    printf("%s  ", "   Coeff.  ");
    printf("%s ", "Intercept ");
    for (i = 1; i <= pl; i++)
      printf("%10" NAG_IFMT "%s", i, i % 4 ? " " : "\n");
    printf("\n");
    if (pl < ip1) {
      for (i = pl + 1; i <= ip1; i++)
        printf("%10" NAG_IFMT "%s", i, i % 5 ? " " : "\n");
      printf("\n");
    }
    pl = MIN(ip + 1, 5);
    for (i = 1; i <= lh; i++) {
      printf("%10.4f", h[i - 1]);
      for (j = 1; j <= pl; j++)
        printf("%10.4f%s", B(j, i), j % 5 ? " " : "\n");
      printf("\n");
      if (pl < ip) {
        for (j = pl + 1; j <= ip; j++)
          printf("%10.4f%s", B(j, i), j % 5 ? " " : "\n");
        printf("\n");
      }
    }
  }
  /* Variance inflation factors */
  if (wantvf != Nag_NoVIF) {
    printf("\n");
    printf("%s\n", "Variance Inflation Factors");
    pl = MIN(ip, 5);
    printf("%s\n", "  Ridge  ");
    printf("%s", "  Coeff.  ");
    for (i = 1; i <= pl; i++)
      printf("%10" NAG_IFMT "%s", i, i % 5 ? " " : "\n");
    printf("\n");
    if (pl < ip) {
      for (i = pl + 1; i <= ip; i++)
        printf("%10" NAG_IFMT "%s", i, i % 5 ? " " : "\n");
      printf("\n");
    }
    for (i = 1; i <= lh; i++) {
      printf("%10.4f", h[i - 1]);
      for (j = 1; j <= pl; j++)
        printf("%10.4f%s", VF(j, i), j % 5 ? " " : "\n");
      printf("\n");
      if (pl < ip) {
        for (j = pl + 1; j <= ip; j++)
          printf("%10.4f%s", VF(j, i), j % 5 ? " " : "\n");
        printf("\n");
      }
    }
  }
  /* Prediction error criterion */
  if (lpec > 0) {
    printf("\n");
    printf("%s\n", "Prediction error criterion");
    pl = MIN(lpec, 5);
    printf("%s\n", "  Ridge  ");
    printf("%s", "  Coeff.  ");
    for (i = 1; i <= pl; i++)
      printf("%10" NAG_IFMT "%s", i, i % 5 ? " " : "\n");
    printf("\n");
    if (pl < lpec) {
      for (i = pl + 1; i <= lpec; i++)
        printf("%10" NAG_IFMT "%s", i, i % 5 ? " " : "\n");
      printf("\n");
    }
    for (i = 1; i <= lh; i++) {
      printf("%10.4f", h[i - 1]);
      for (j = 1; j <= pl; j++)
        printf("%10.4f%s", PE(j, i), j % 5 ? " " : "\n");
      if (pl < ip) {
        for (j = pl + 1; j <= ip; j++)
          printf("%10.4f%s", PE(j, i), j % 5 ? " " : "\n");
      }
    }
    printf("\n");
    printf("%s\n", "Key:");
    for (i = 1; i <= lpec; i++) {
      if (pec[i - 1] == Nag_LOOCV) {
        printf("  %5" NAG_IFMT "  Leave one out cross-validation\n", i);
      } else if (pec[i - 1] == Nag_GCV) {
        printf("  %5" NAG_IFMT "  Generalized cross-validation\n", i);
      } else if (pec[i - 1] == Nag_EUV) {
        printf("  %5" NAG_IFMT "  Unbiased estimate of variance\n", i);
      } else if (pec[i - 1] == Nag_FPE) {
        printf("  %5" NAG_IFMT "  Final prediction error\n", i);
      } else if (pec[i - 1] == Nag_BIC) {
        printf("  %5" NAG_IFMT "  Bayesian information criterion\n", i);
      }
    }
  }

END:
  NAG_FREE(b);
  NAG_FREE(h);
  NAG_FREE(nep);
  NAG_FREE(pe);
  NAG_FREE(vf);
  NAG_FREE(x);
  NAG_FREE(y);
  NAG_FREE(isx);
  NAG_FREE(pec);

  return exit_status;
}