X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Fmath%2Flinreg%2Flinreg.c;h=0782c36fc3df82aae30d00c5f88fbd5f89aa3ad7;hb=5b5f08a99c564c2f3e981ae0efe21ffc3c578ac4;hp=9ba84f60ece9d8c3991be31cce09766129e886ef;hpb=43b1296aafe7582e7dbe6c2b6a8b478d7d9b0fcf;p=pspp-builds.git

diff --git a/src/math/linreg/linreg.c b/src/math/linreg/linreg.c
index 9ba84f60..0782c36f 100644
--- a/src/math/linreg/linreg.c
+++ b/src/math/linreg/linreg.c
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 2005 Free Software Foundation, Inc. Written by Jason H. Stover.
+   Copyright (C) 2005 Free Software Foundation, Inc. 
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -94,9 +94,9 @@ int
 pspp_linreg_get_vars (const void *c_, const struct variable **v)
 {
   const pspp_linreg_cache *c = c_;
-  struct pspp_coeff *coef = NULL;
   const struct variable *tmp;
   int i;
+  int j;
   int result = 0;
 
   /*
@@ -107,15 +107,9 @@ pspp_linreg_get_vars (const void *c_, const struct variable **v)
     {
       v[i] = NULL;
     }
-  /*
-     Start at c->coeff[1] to avoid the intercept.
-   */
-  v[result] =  pspp_coeff_get_var (c->coeff[1], 0);
-  result = (v[result] == NULL) ? 0 : 1;
-
-  for (coef = c->coeff[2]; coef < c->coeff[c->n_coeffs]; coef++)
+  for (j = 0; j < c->n_coeffs; j++)
     {
-      tmp = pspp_coeff_get_var (coef, 0);
+      tmp = pspp_coeff_get_var (c->coeff[j], 0);
       assert (tmp != NULL);
       /* Repeated variables are likely to bunch together, at the end
          of the array. */
@@ -184,10 +178,12 @@ pspp_linreg_cache_free (void *m)
       gsl_vector_free (c->indep_std);
       gsl_vector_free (c->ss_indeps);
       gsl_matrix_free (c->cov);
+      gsl_vector_free (c->ssx);
       for (i = 0; i < c->n_coeffs; i++)
 	{
 	  pspp_coeff_free (c->coeff[i]);
 	}
+      free (c->coeff);
       free (c);
     }
   return true;
@@ -203,7 +199,7 @@ pspp_linreg (const gsl_vector * Y, const gsl_matrix * X,
 	     const pspp_linreg_opts * opts, pspp_linreg_cache * cache)
 {
   int rc;
-  gsl_matrix *design;
+  gsl_matrix *design = NULL;
   gsl_matrix_view xtx;
   gsl_matrix_view xm;
   gsl_matrix_view xmxtx;
@@ -244,9 +240,9 @@ pspp_linreg (const gsl_vector * Y, const gsl_matrix * X,
   cache->dft = cache->n_obs - 1;
   cache->dfm = cache->n_indeps;
   cache->dfe = cache->dft - cache->dfm;
-  cache->n_coeffs = X->size2 + 1;	/* Adjust this later to allow for
-					   regression through the origin.
-					 */
+  cache->n_coeffs = X->size2;
+  cache->intercept = 0.0;
+
   if (cache->method == PSPP_LINREG_SWEEP)
     {
       gsl_matrix *sw;
@@ -320,7 +316,7 @@ pspp_linreg (const gsl_vector * Y, const gsl_matrix * X,
       for (i = 0; i < cache->n_indeps; i++)
 	{
 	  tmp = gsl_matrix_get (sw, i, cache->n_indeps);
-	  cache->coeff[i + 1]->estimate = tmp;
+	  cache->coeff[i]->estimate = tmp;
 	  m -= tmp * gsl_vector_get (cache->indep_means, i);
 	}
       /*
@@ -356,7 +352,7 @@ pspp_linreg (const gsl_vector * Y, const gsl_matrix * X,
 	    }
 	  gsl_matrix_set (cache->cov, 0, 0, tmp);
 
-	  cache->coeff[0]->estimate = m;
+	  cache->intercept = m;
 	}
       else
 	{
@@ -366,6 +362,18 @@ pspp_linreg (const gsl_vector * Y, const gsl_matrix * X,
 	}
       gsl_matrix_free (sw);
     }
+  else if (cache->method == PSPP_LINREG_CONDITIONAL_INVERSE)
+    {
+      /*
+	Use the SVD of X^T X to find a conditional inverse of X^TX. If
+	the SVD is X^T X = U D V^T, then set the conditional inverse
+	to (X^T X)^c = V D^- U^T. D^- is defined as follows: If entry
+	(i, i) has value sigma_i, then entry (i, i) of D^- is 1 /
+	sigma_i if sigma_i > 0, and 0 otherwise. Then solve the normal
+	equations by setting the estimated parameter vector to 
+	(X^TX)^c X^T Y.
+       */
+    }
   else
     {
       gsl_multifit_linear_workspace *wk;
@@ -389,10 +397,11 @@ pspp_linreg (const gsl_vector * Y, const gsl_matrix * X,
       wk = gsl_multifit_linear_alloc (design->size1, design->size2);
       rc = gsl_multifit_linear (design, Y, param_estimates,
 				cache->cov, &(cache->sse), wk);
-      for (i = 0; i < cache->n_coeffs; i++)
+      for (i = 1; i < cache->n_coeffs; i++)
 	{
 	  cache->coeff[i]->estimate = gsl_vector_get (param_estimates, i);
 	}
+      cache->intercept = gsl_vector_get (param_estimates, 0);
       if (rc == GSL_SUCCESS)
 	{
 	  gsl_multifit_linear_free (wk);