-/* PSPP - computes sample statistics.
- Copyright (C) 2006 Free Software Foundation, Inc.
+/* PSPP - a program for statistical analysis.
+ Copyright (C) 2006, 2009 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 the Free Software Foundation; either version 2 of the
- License, or (at your option) any later version.
+ 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
+ the Free Software Foundation, either version 3 of the License, or
+ (at your option) any later version.
- This program is distributed in the hope that it will be useful, but
- WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- General Public License for more details.
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
You should have received a copy of the GNU General Public License
- along with this program; if not, write to the Free Software
- Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
- 02110-1301, USA. */
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
#include <config.h>
#include <libpspp/compiler.h>
#include <output/table.h>
-#include <libpspp/alloc.h>
+#include <data/format.h>
#include <data/case.h>
#include <data/casereader.h>
#include <data/dictionary.h>
#include "binomial.h"
#include "freq.h"
+#include "xalloc.h"
+
#include "gettext.h"
#define _(msgid) gettext (msgid)
#include <gsl/gsl_cdf.h>
#include <gsl/gsl_randist.h>
-#include <gsl-extras/gsl-extras.h>
#include <minmax.h>
static double
calculate_binomial_internal (double n1, double n2, double p)
{
- /* SPSS Statistical Algorithms has completely different and WRONG
+ /* SPSS Statistical Algorithms has completely different and WRONG
advice here. */
- double sig1tailed = gslextras_cdf_binomial_P (n1, n1 + n2, p);
+ double sig1tailed = gsl_cdf_binomial_P (n1, p, n1 + n2);
if ( p == 0.5 )
return sig1tailed > 0.5 ? 1.0 :sig1tailed * 2.0;
bool warn = true;
const struct one_sample_test *ost = (const struct one_sample_test *) bst;
- struct ccase c;
+ struct ccase *c;
- while (casereader_read(input, &c))
+ while ((c = casereader_read(input)) != NULL)
{
int v;
- double w = dict_get_case_weight (dict, &c, &warn);
+ double w = dict_get_case_weight (dict, c, &warn);
for (v = 0 ; v < ost->n_vars ; ++v )
{
const struct variable *var = ost->vars[v];
- const union value *value = case_data (&c, var);
+ const union value *value = case_data (c, var);
int width = var_get_width (var);
if (var_is_value_missing (var, value, exclude))
- break;
+ continue;
- if ( NULL == cat1[v].value )
+ if (bst->cutpoint != SYSMIS)
{
- cat1[v].value = value_dup (value, width);
- cat1[v].count = w;
+ if ( compare_values_short (cat1[v].value, value, var) >= 0 )
+ cat1[v].count += w;
+ else
+ cat2[v].count += w;
}
- else if ( 0 == compare_values (cat1[v].value, value, width))
- cat1[v].count += w;
- else if ( NULL == cat2[v].value )
+ else
{
- cat2[v].value = value_dup (value, width);
- cat2[v].count = w;
+ if ( NULL == cat1[v].value )
+ {
+ cat1[v].value = value_dup (value, width);
+ cat1[v].count = w;
+ }
+ else if ( 0 == compare_values_short (cat1[v].value, value, var))
+ cat1[v].count += w;
+ else if ( NULL == cat2[v].value )
+ {
+ cat2[v].value = value_dup (value, width);
+ cat2[v].count = w;
+ }
+ else if ( 0 == compare_values_short (cat2[v].value, value, var))
+ cat2[v].count += w;
+ else if ( bst->category1 == SYSMIS)
+ msg (ME, _("Variable %s is not dichotomous"), var_get_name (var));
}
- else if ( 0 == compare_values (cat2[v].value, value, width))
- cat2[v].count += w;
- else if ( bst->category1 == SYSMIS)
- msg (ME, _("Variable %s is not dichotomous"), var_get_name (var));
}
- case_destroy (&c);
+ case_unref (c);
}
return casereader_destroy (input);
}
binomial_execute (const struct dataset *ds,
struct casereader *input,
enum mv_class exclude,
- const struct npar_test *test)
+ const struct npar_test *test,
+ bool exact UNUSED,
+ double timer UNUSED)
{
int v;
+ const struct dictionary *dict = dataset_dict (ds);
const struct binomial_test *bst = (const struct binomial_test *) test;
const struct one_sample_test *ost = (const struct one_sample_test*) test;
struct freq_mutable *cat1 = xzalloc (sizeof (*cat1) * ost->n_vars);
struct freq_mutable *cat2 = xzalloc (sizeof (*cat1) * ost->n_vars);
- assert ((bst->category1 == SYSMIS) == (bst->category2 == SYSMIS) );
+ assert ((bst->category1 == SYSMIS) == (bst->category2 == SYSMIS) || bst->cutpoint != SYSMIS);
- if ( bst->category1 != SYSMIS )
+ if ( bst->cutpoint != SYSMIS )
+ {
+ int i;
+ union value v;
+ v.f = bst->cutpoint;
+ for (i = 0; i < ost->n_vars; i++)
+ cat1[i].value = value_dup (&v, 0);
+ }
+ else if ( bst->category1 != SYSMIS )
{
+ int i;
union value v;
v.f = bst->category1;
- cat1->value = value_dup (&v, 0);
+ for (i = 0; i < ost->n_vars; i++)
+ cat1[i].value = value_dup (&v, 0);
}
if ( bst->category2 != SYSMIS )
{
+ int i;
union value v;
v.f = bst->category2;
- cat2->value = value_dup (&v, 0);
+ for (i = 0; i < ost->n_vars; i++)
+ cat2[i].value = value_dup (&v, 0);
}
- if (do_binomial (dataset_dict(ds), input, bst, cat1, cat2, exclude))
+ if (do_binomial (dict, input, bst, cat1, cat2, exclude))
{
+ const struct variable *wvar = dict_get_weight (dict);
+ const struct fmt_spec *wfmt = wvar ?
+ var_get_print_format (wvar) : & F_8_0;
+
struct tab_table *table = tab_create (7, ost->n_vars * 3 + 1, 0);
tab_dim (table, tab_natural_dimensions);
for (v = 0 ; v < ost->n_vars; ++v)
{
double n_total, sig;
+ struct string catstr1;
+ struct string catstr2;
const struct variable *var = ost->vars[v];
+
+ ds_init_empty (&catstr1);
+ ds_init_empty (&catstr2);
+
+ if ( bst->cutpoint != SYSMIS)
+ {
+ ds_put_format (&catstr1, "<= %g", bst->cutpoint);
+ }
+ else
+ {
+ var_append_value_name (var, cat1[v].value, &catstr1);
+ var_append_value_name (var, cat2[v].value, &catstr2);
+ }
+
tab_hline (table, TAL_1, 0, tab_nc (table) -1, 1 + v * 3);
/* Titles */
tab_text (table, 1, 3 + v * 3, TAB_LEFT, _("Total"));
/* Test Prop */
- tab_float (table, 5, 1 + v * 3, TAB_NONE, bst->p, 8, 3);
+ tab_double (table, 5, 1 + v * 3, TAB_NONE, bst->p, NULL);
/* Category labels */
- tab_text (table, 2, 1 + v * 3, TAB_NONE,
- var_get_value_name (var, cat1[v].value));
- tab_text (table, 2, 2 + v * 3, TAB_NONE,
- var_get_value_name (var, cat2[v].value));
+ tab_text (table, 2, 1 + v * 3, TAB_NONE, ds_cstr (&catstr1));
+ tab_text (table, 2, 2 + v * 3, TAB_NONE, ds_cstr (&catstr2));
/* Observed N */
- tab_float (table, 3, 1 + v * 3, TAB_NONE, cat1[v].count, 8, 0);
- tab_float (table, 3, 2 + v * 3, TAB_NONE, cat2[v].count, 8, 0);
+ tab_double (table, 3, 1 + v * 3, TAB_NONE, cat1[v].count, wfmt);
+ tab_double (table, 3, 2 + v * 3, TAB_NONE, cat2[v].count, wfmt);
n_total = cat1[v].count + cat2[v].count;
- tab_float (table, 3, 3 + v * 3, TAB_NONE, n_total, 8, 0);
+ tab_double (table, 3, 3 + v * 3, TAB_NONE, n_total, wfmt);
/* Observed Proportions */
- tab_float (table, 4, 1 + v * 3, TAB_NONE,
- cat1[v].count / n_total, 8, 3);
- tab_float (table, 4, 2 + v * 3, TAB_NONE,
- cat2[v].count / n_total, 8, 3);
- tab_float (table, 4, 3 + v * 3, TAB_NONE,
- (cat1[v].count + cat2[v].count) / n_total, 8, 2);
+ tab_double (table, 4, 1 + v * 3, TAB_NONE,
+ cat1[v].count / n_total, NULL);
+ tab_double (table, 4, 2 + v * 3, TAB_NONE,
+ cat2[v].count / n_total, NULL);
+
+ tab_double (table, 4, 3 + v * 3, TAB_NONE,
+ (cat1[v].count + cat2[v].count) / n_total, NULL);
/* Significance */
sig = calculate_binomial (cat1[v].count, cat2[v].count, bst->p);
- tab_float (table, 6, 1 + v * 3, TAB_NONE, sig, 8, 3);
+ tab_double (table, 6, 1 + v * 3, TAB_NONE, sig, NULL);
+
+ ds_destroy (&catstr1);
+ ds_destroy (&catstr2);
}
tab_text (table, 2, 0, TAB_CENTER, _("Category"));
tab_vline (table, TAL_2, 2, 0, tab_nr (table) -1);
tab_submit (table);
}
-
- for (v = 0; v < ost->n_vars; v++)
+
+ for (v = 0; v < ost->n_vars; v++)
{
free (cat1[v].value);
- free (cat2[v].value);
+ free (cat2[v].value);
}
free (cat1);
- free (cat2);
+ free (cat2);
}