1 /* Pspp - a program for statistical analysis.
2 Copyright (C) 2010, 2011 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>. */
20 #include "kruskal-wallis.h"
22 #include <gsl/gsl_cdf.h>
25 #include "data/casereader.h"
26 #include "data/casewriter.h"
27 #include "data/dataset.h"
28 #include "data/dictionary.h"
29 #include "data/format.h"
30 #include "data/subcase.h"
31 #include "data/variable.h"
32 #include "libpspp/assertion.h"
33 #include "libpspp/hmap.h"
34 #include "libpspp/message.h"
35 #include "libpspp/misc.h"
36 #include "math/sort.h"
37 #include "output/tab.h"
39 #include "gl/minmax.h"
40 #include "gl/xalloc.h"
43 /* Returns true iff the independent variable lies in the range [nst->val1, nst->val2] */
45 include_func (const struct ccase *c, void *aux)
47 const struct n_sample_test *nst = aux;
49 if (0 < value_compare_3way (&nst->val1, case_data (c, nst->indep_var), var_get_width (nst->indep_var)))
52 if (0 > value_compare_3way (&nst->val2, case_data (c, nst->indep_var), var_get_width (nst->indep_var)))
61 struct hmap_node node;
68 /* Return the entry with the key GROUP or null if there is no such entry */
69 static struct rank_entry *
70 find_rank_entry (const struct hmap *map, const union value *group, size_t width)
72 struct rank_entry *re = NULL;
73 size_t hash = value_hash (group, width, 0);
75 HMAP_FOR_EACH_WITH_HASH (re, struct rank_entry, node, hash, map)
77 if (0 == value_compare_3way (group, &re->group, width))
84 /* Calculates the adjustment necessary for tie compensation */
86 distinct_callback (double v UNUSED, casenumber t, double w UNUSED, void *aux)
88 double *tiebreaker = aux;
90 *tiebreaker += pow3 (t) - t;
100 static void show_ranks_box (const struct n_sample_test *nst, const struct kw *kw, int n_groups);
101 static void show_sig_box (const struct n_sample_test *nst, const struct kw *kw);
104 kruskal_wallis_execute (const struct dataset *ds,
105 struct casereader *input,
106 enum mv_class exclude,
107 const struct npar_test *test,
114 const struct dictionary *dict = dataset_dict (ds);
115 const struct n_sample_test *nst = UP_CAST (test, const struct n_sample_test, parent);
116 const struct caseproto *proto ;
119 int total_n_groups = 0.0;
121 struct kw *kw = xcalloc (nst->n_vars, sizeof *kw);
123 /* If the independent variable is missing, then we ignore the case */
124 input = casereader_create_filter_missing (input,
129 input = casereader_create_filter_weight (input, dict, &warn, NULL);
131 /* Remove all those cases which are outside the range (val1, val2) */
132 input = casereader_create_filter_func (input, include_func, NULL,
133 CONST_CAST (struct n_sample_test *, nst), NULL);
135 proto = casereader_get_proto (input);
136 rank_idx = caseproto_get_n_widths (proto);
138 /* Rank cases by the v value */
139 for (i = 0; i < nst->n_vars; ++i)
141 double tiebreaker = 0.0;
143 enum rank_error rerr = 0;
144 struct casereader *rr;
145 struct casereader *r = casereader_clone (input);
147 r = sort_execute_1var (r, nst->vars[i]);
149 /* Ignore missings in the test variable */
150 r = casereader_create_filter_missing (r, &nst->vars[i], 1,
154 rr = casereader_create_append_rank (r,
156 dict_get_weight (dict),
158 distinct_callback, &tiebreaker);
160 hmap_init (&kw[i].map);
161 for (; (c = casereader_read (rr)); case_unref (c))
163 const union value *group = case_data (c, nst->indep_var);
164 const size_t group_var_width = var_get_width (nst->indep_var);
165 struct rank_entry *rank = find_rank_entry (&kw[i].map, group, group_var_width);
169 rank = xzalloc (sizeof *rank);
170 value_clone (&rank->group, group, group_var_width);
172 hmap_insert (&kw[i].map, &rank->node,
173 value_hash (&rank->group, group_var_width, 0));
176 rank->sum_of_ranks += case_data_idx (c, rank_idx)->f;
177 rank->n += dict_get_case_weight (dict, c, &warn);
179 /* If this assertion fires, then either the data wasn't sorted or some other
184 casereader_destroy (rr);
186 /* Calculate the value of h */
188 struct rank_entry *mre;
191 HMAP_FOR_EACH (mre, struct rank_entry, node, &kw[i].map)
193 kw[i].h += pow2 (mre->sum_of_ranks) / mre->n;
198 kw[i].h *= 12 / (n * ( n + 1));
199 kw[i].h -= 3 * (n + 1) ;
201 kw[i].h /= 1 - tiebreaker/ (pow3 (n) - n);
205 casereader_destroy (input);
207 show_ranks_box (nst, kw, total_n_groups);
208 show_sig_box (nst, kw);
210 /* Cleanup allocated memory */
211 for (i = 0 ; i < nst->n_vars; ++i)
213 struct rank_entry *mre, *next;
214 HMAP_FOR_EACH_SAFE (mre, next, struct rank_entry, node, &kw[i].map)
216 hmap_delete (&kw[i].map, &mre->node);
219 hmap_destroy (&kw[i].map);
227 #define _(msgid) gettext (msgid)
231 show_ranks_box (const struct n_sample_test *nst, const struct kw *kw, int n_groups)
235 const int row_headers = 2;
236 const int column_headers = 1;
237 struct tab_table *table =
238 tab_create (row_headers + 2, column_headers + n_groups + nst->n_vars);
240 tab_headers (table, row_headers, 0, column_headers, 0);
242 tab_title (table, _("Ranks"));
244 /* Vertical lines inside the box */
245 tab_box (table, 1, 0, -1, TAL_1,
246 row_headers, 0, tab_nc (table) - 1, tab_nr (table) - 1 );
248 /* Box around the table */
249 tab_box (table, TAL_2, TAL_2, -1, -1,
250 0, 0, tab_nc (table) - 1, tab_nr (table) - 1 );
252 tab_text (table, 1, 0, TAT_TITLE,
253 var_to_string (nst->indep_var)
256 tab_text (table, 3, 0, 0, _("Mean Rank"));
257 tab_text (table, 2, 0, 0, _("N"));
259 tab_hline (table, TAL_2, 0, tab_nc (table) -1, column_headers);
260 tab_vline (table, TAL_2, row_headers, 0, tab_nr (table) - 1);
263 row = column_headers;
264 for (i = 0 ; i < nst->n_vars ; ++i)
267 const struct rank_entry *re;
270 tab_hline (table, TAL_1, 0, tab_nc (table) -1, row);
272 tab_text (table, 0, row,
273 TAT_TITLE, var_to_string (nst->vars[i]));
275 HMAP_FOR_EACH (re, const struct rank_entry, node, &kw[i].map)
278 ds_init_empty (&str);
280 var_append_value_name (nst->indep_var, &re->group, &str);
282 tab_text (table, 1, row, TAB_LEFT, ds_cstr (&str));
283 tab_double (table, 2, row, TAB_LEFT, re->n, &F_8_0);
284 tab_double (table, 3, row, TAB_LEFT, re->sum_of_ranks / re->n, 0);
290 tab_double (table, 2, row, TAB_LEFT,
292 tab_text (table, 1, row++, TAB_LEFT, _("Total"));
300 show_sig_box (const struct n_sample_test *nst, const struct kw *kw)
303 const int row_headers = 1;
304 const int column_headers = 1;
305 struct tab_table *table =
306 tab_create (row_headers + nst->n_vars * 2, column_headers + 3);
308 tab_headers (table, row_headers, 0, column_headers, 0);
310 tab_title (table, _("Test Statistics"));
312 tab_text (table, 0, column_headers,
313 TAT_TITLE | TAB_LEFT , _("Chi-Square"));
315 tab_text (table, 0, 1 + column_headers,
316 TAT_TITLE | TAB_LEFT, _("df"));
318 tab_text (table, 0, 2 + column_headers,
319 TAT_TITLE | TAB_LEFT, _("Asymp. Sig."));
321 /* Box around the table */
322 tab_box (table, TAL_2, TAL_2, -1, -1,
323 0, 0, tab_nc (table) - 1, tab_nr (table) - 1 );
326 tab_hline (table, TAL_2, 0, tab_nc (table) -1, column_headers);
327 tab_vline (table, TAL_2, row_headers, 0, tab_nr (table) - 1);
329 for (i = 0 ; i < nst->n_vars; ++i)
331 const double df = hmap_count (&kw[i].map) - 1;
332 tab_text (table, column_headers + 1 + i, 0, TAT_TITLE,
333 var_to_string (nst->vars[i])
336 tab_double (table, column_headers + 1 + i, 1, 0,
339 tab_double (table, column_headers + 1 + i, 2, 0,
342 tab_double (table, column_headers + 1 + i, 3, 0,
343 gsl_cdf_chisq_Q (kw[i].h, df),