/* PSPP - a program for statistical analysis.
- Copyright (C) 2010 Free Software Foundation, Inc.
+ Copyright (C) 2010, 2011 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
#include <config.h>
-#include "mann-whitney.h"
+#include "language/stats/mann-whitney.h"
#include <gsl/gsl_cdf.h>
-#include <libpspp/cast.h>
-#include <libpspp/misc.h>
-
-#include <math/sort.h>
-#include <data/case.h>
-#include <data/dictionary.h>
-#include <data/procedure.h>
-#include <data/casereader.h>
-#include <data/format.h>
-#include <data/variable.h>
+#include "data/case.h"
+#include "data/casereader.h"
+#include "data/dataset.h"
+#include "data/dictionary.h"
+#include "data/format.h"
+#include "data/variable.h"
+#include "libpspp/cast.h"
+#include "libpspp/misc.h"
+#include "math/sort.h"
+#include "output/tab.h"
/* Calculates the adjustment necessary for tie compensation */
static void
double u; /* The Mann-Whitney U statistic */
double w; /* The Wilcoxon Rank Sum W statistic */
- double z;
+ double z;
};
static void show_ranks_box (const struct n_sample_test *nst, const struct mw *mw);
static void show_statistics_box (const struct n_sample_test *nst, const struct mw *mw, bool exact);
+
+static bool
+belongs_to_test (const struct ccase *c, void *aux)
+{
+ const struct n_sample_test *nst = aux;
+
+ const union value *group = case_data (c, nst->indep_var);
+ const size_t group_var_width = var_get_width (nst->indep_var);
+
+ if ( value_equal (group, &nst->val1, group_var_width))
+ return true;
+
+ if ( value_equal (group, &nst->val2, group_var_width))
+ return true;
+
+ return false;
+}
+
+
+
void
mann_whitney_execute (const struct dataset *ds,
struct casereader *input,
enum mv_class exclude,
const struct npar_test *test,
bool exact,
- double timer)
+ double timer UNUSED)
{
int i;
const struct dictionary *dict = dataset_dict (ds);
struct casereader *rr;
struct ccase *c;
const struct variable *var = nst->vars[i];
-
+
struct casereader *reader =
- sort_execute_1var (casereader_clone (input), var);
+ casereader_create_filter_func (casereader_clone (input),
+ belongs_to_test,
+ NULL,
+ CONST_CAST (struct n_sample_test *, nst),
+ NULL);
+
+ reader = casereader_create_filter_missing (reader, &var, 1,
+ exclude,
+ NULL, NULL);
+
+ reader = sort_execute_1var (reader, var);
rr = casereader_create_append_rank (reader, var,
dict_get_weight (dict),
for (; (c = casereader_read (rr)); case_unref (c))
{
- const union value *val = case_data (c, var);
const union value *group = case_data (c, nst->indep_var);
const size_t group_var_width = var_get_width (nst->indep_var);
const double rank = case_data_idx (c, rank_idx)->f;
- if ( var_is_value_missing (var, val, exclude))
- continue;
-
if ( value_equal (group, &nst->val1, group_var_width))
{
mw[i].rank_sum[0] += rank;
denominator -= tiebreaker;
denominator *= mwv->n[0] * mwv->n[1];
denominator /= n * (n - 1);
-
+
mwv->z /= sqrt (denominator);
}
}
\f
-#include <output/tab.h>
#include "gettext.h"
#define _(msgid) gettext (msgid)
var_to_string (nst->vars[i]));
tab_double (table, 1, column_headers + i, 0,
- mw->n[0], 0);
+ mw->n[0], NULL, RC_OTHER);
tab_double (table, 2, column_headers + i, 0,
- mw->n[1], 0);
+ mw->n[1], NULL, RC_OTHER);
tab_double (table, 3, column_headers + i, 0,
- mw->n[1] + mw->n[0], 0);
+ mw->n[1] + mw->n[0], NULL, RC_OTHER);
/* Mean Ranks */
tab_double (table, 4, column_headers + i, 0,
- mw->rank_sum[0] / mw->n[0], 0);
+ mw->rank_sum[0] / mw->n[0], NULL, RC_OTHER);
tab_double (table, 5, column_headers + i, 0,
- mw->rank_sum[1] / mw->n[1], 0);
+ mw->rank_sum[1] / mw->n[1], NULL, RC_OTHER);
/* Sum of Ranks */
tab_double (table, 6, column_headers + i, 0,
- mw->rank_sum[0], 0);
+ mw->rank_sum[0], NULL, RC_OTHER);
tab_double (table, 7, column_headers + i, 0,
- mw->rank_sum[1], 0);
+ mw->rank_sum[1], NULL, RC_OTHER);
}
tab_submit (table);
tab_text (table, 3, 0, TAT_TITLE | TAB_CENTER, _("Z"));
tab_text (table, 4, 0, TAT_TITLE | TAB_CENTER, _("Asymp. Sig. (2-tailed)"));
- if (exact)
+ if (exact)
{
tab_text (table, 5, 0, TAT_TITLE | TAB_CENTER, _("Exact Sig. (2-tailed)"));
tab_text (table, 6, 0, TAT_TITLE | TAB_CENTER, _("Point Probability"));
var_to_string (nst->vars[i]));
tab_double (table, 1, column_headers + i, 0,
- mw->u, 0);
+ mw->u, NULL, RC_OTHER);
tab_double (table, 2, column_headers + i, 0,
- mw->w, 0);
+ mw->w, NULL, RC_OTHER);
tab_double (table, 3, column_headers + i, 0,
- mw->z, 0);
+ mw->z, NULL, RC_OTHER);
tab_double (table, 4, column_headers + i, 0,
- 2.0 * gsl_cdf_ugaussian_P (mw->z), 0);
+ 2.0 * gsl_cdf_ugaussian_P (mw->z), NULL, RC_PVALUE);
}
tab_submit (table);