+++ /dev/null
-#include <config.h>
-#include "wx-mp-sr.h"
-
-/*********************************************************************
-*
-* Calculate the exact level of significance for a
-* Wilcoxon Matched-Pair Signed-Ranks Test using the sample's
-* Sum of Ranks W and the sample size (i.e., number of pairs) N.
-* This whole routine can be run as a stand-alone program.
-*
-* Use:
-* WX-MP-SR W N
-*
-* Copyright 1996, Rob van Son
-*
-* 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 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.
-*
-* -------------------------------------------------------
-* Rob van Son
-* Institute of Phonetic Sciences & IFOTT
-* University of Amsterdam, Spuistraat 210
-* NL-1012VT Amsterdam, The Netherlands
-* Tel.: (+31) 205252196 Fax.: (+31) 205252197
-* Email: r.j.j.h.vanson@uva.nl
-* WWW page: http://www.fon.hum.uva.nl/rob
-* -------------------------------------------------------
-*
-* This is the actual routine that calculates the exact (two-tailed)
-* level of significance for the Wilcoxon Matched-Pairs Signed-Ranks
-* test. The inputs are the Sum of Ranks of either the positive of
-* negative samples (W) and the sample size (N).
-* The Level of significance is calculated by checking for each
-* possible outcome (2**N possibilities) whether the sum of ranks
-* is larger than or equal to the observed Sum of Ranks (W).
-*
-* NOTE: The execution-time scales like ~ N*2**N, i.e., N*pow(2, N),
-* which is more than exponential. Adding a single pair to the sample
-* (i.e., increase N by 1) will more than double the time needed to
-* complete the calculations (apart from an additive constant).
-* The execution-time of this program can easily outrun your
-* patience.
-*
-***********************************************************************/
-
-double LevelOfSignificanceWXMPSR(double Winput, long int N)
-{
- unsigned long int W, MaximalW, NumberOfPossibilities, CountLarger;
- unsigned long int i, RankSum, j;
- double p;
-
- /* Determine Wmax, i.e., work with the largest Rank Sum */
- MaximalW = N*(N+1)/2;
- if(Winput < MaximalW/2)Winput = MaximalW - Winput;
- W = Winput; /* Convert to long int */
- if(W != Winput)++W; /* Increase to next full integer */
-
- /* The total number of possible outcomes is 2**N */
- NumberOfPossibilities = 1 << N;
-
- /* Initialize and loop. The loop-interior will be run 2**N times. */
- CountLarger = 0;
- /* Generate all distributions of sign over ranks as bit-patterns (i). */
- for(i=0; i < NumberOfPossibilities; ++i)
- {
- RankSum = 0;
- /*
- Shift "sign" bits out of i to determine the Sum of Ranks (j).
- */
- for(j=0; j < N; ++j)
- {
- if((i >> j) & 1)RankSum += j + 1;
- };
- /*
- * Count the number of "samples" that have a Sum of Ranks larger than
- * or equal to the one found (i.e., >= W).
- */
- if(RankSum >= W)++CountLarger;
- };
- /*****************************************************************
- * The level of significance is the number of outcomes with a
- * sum of ranks equal to or larger than the one found (W)
- * divided by the total number of possible outcomes.
- * The level is doubled to get the two-tailed result.
- ******************************************************************/
- p = 2*((double)CountLarger) / ((double)NumberOfPossibilities);
-
- return p;
-}
-
--- /dev/null
+/* PSPP - a program for statistical analysis.
+ Copyright (C) 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 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.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>. */
+
+/* Thanks to Rob van Son for writing the original version of this
+ file. This version has been completely rewritten; only the
+ name of the function has been retained. In the process of
+ rewriting, it was sped up from O(2**N) to O(N**3). */
+
+#include <config.h>
+#include "wilcoxon-sig.h"
+#include <assert.h>
+#include <limits.h>
+#include <math.h>
+#include <stdlib.h>
+#include "xalloc.h"
+
+/* For integers N and W, with 0 <= N < CHAR_BIT*sizeof(long),
+ calculates and returns the value of the function S(N,W),
+ defined as the number of subsets of 1, 2, 3, ..., N that sum
+ to at least W. There are 2**N subsets of N items, so S(N,W)
+ is in the range 0...2**N.
+
+ There are a few trivial cases:
+
+ * For W <= 0, S(N,W) = 2**N.
+
+ * For W > N*(N+1)/2, S(N,W) = 0.
+
+ * S(1,1) = 1.
+
+ Notably, these trivial cases include all values of W for N = 1.
+
+ Now consider the remaining, nontrivial cases, that is, N > 1 and
+ 1 <= W <= N*(N+1)/2. In this case, apply the following identity:
+
+ S(N,W) = S(N-1, W) + S(N-1, W-N).
+
+ The first term on the right hand is the number of subsets that do
+ not include N that sum to at least W; the second term is the
+ number of subsets that do include N that sum to at least W.
+
+ Then we repeatedly apply the identity to the result, reducing the
+ value of N by 1 each time until we reach N=1. Some expansions
+ yield trivial cases, e.g. if W - N <= 0 (in which case we add a
+ 2**N term to the final result) or if W is greater than the new N.
+
+ Here is an example:
+
+ S(7,7) = S(6,7) + S(6,0)
+ = S(6,7) + 64
+
+ = (S(5,7) + S(5,1)) + 64
+
+ = (S(4,7) + S(4,2)) + (S(4,1) + S(4,0)) + 64
+ = S(4,7) + S(4,2) + S(4,1) + 80
+
+ = (S(3,7) + S(3,3)) + (S(3,2) + S(3,2)) + (S(3,1) + S(3,0)) + 80
+ = S(3,3) + 2*S(3,2) + S(3,1) + 88
+
+ = (S(2,3) + S(2,0)) + 2*(S(2,2) + S(2,0)) + (S(2,1) + S(2,0)) + 88
+ = S(2,3) + 2*S(2,2) + S(2,1) + 104
+
+ = (S(1,3) + S(1,1)) + 2*(S(1,2) + S(1,0)) + (S(1,1) + S(2,0)) + 104
+ = 2*S(1,1) + 112
+
+ = 114
+
+ This function runs in O(N*W) = O(N**3) time. It seems very
+ likely that it can be made to run in O(N**2) time or perhaps
+ even better, but N is, practically speaking, quite small.
+ Plus, the return value may be as large as N**2, so N must not
+ be larger than 31 on 32-bit systems anyhow, and even 63**3 is
+ only 250,047.
+*/
+static unsigned long int
+count_sums_to_W (unsigned long int n, unsigned long int w)
+{
+ /* The array contain ints even though everything else is long,
+ but no element in the array can have a value bigger than N,
+ and using int will save some memory on 64-bit systems. */
+ unsigned long int total;
+ unsigned long int max;
+ int *array;
+
+ assert (n < CHAR_BIT * sizeof (unsigned long int));
+ if (n == 0)
+ return 0;
+ else if (w <= 0)
+ return 1 << n;
+ else if (w > n * (n + 1) / 2)
+ return 0;
+ else if (n == 1)
+ return 1;
+
+ array = xcalloc (sizeof *array, w + 1);
+ array[w] = 1;
+
+ max = w;
+ total = 0;
+ for (; n > 1; n--)
+ {
+ unsigned long int max_sum = n * (n + 1) / 2;
+ int i;
+
+ if (max_sum < max)
+ max = max_sum;
+
+ for (i = 1; i <= max; i++)
+ if (array[i] != 0)
+ {
+ int new_w = i - n;
+ if (new_w <= 0)
+ total += array[i] * (1 << (n - 1));
+ else
+ array[new_w] += array[i];
+ }
+ }
+ total += array[1];
+ free (array);
+ return total;
+}
+
+/* Returns the exact, two-tailed level of significance for the
+ Wilcoxon Matched-Pairs Signed-Ranks test, given sum of ranks
+ of positive (or negative samples) W and sample size N.
+
+ Returns -1 if the exact significance level cannot be
+ calculated because W is out of the supported range. */
+double
+LevelOfSignificanceWXMPSR (double w, long int n)
+{
+ unsigned long int max_w;
+
+ /* Limit N to valid range that won't cause integer overflow. */
+ if (n < 0 || n >= CHAR_BIT * sizeof (unsigned long int))
+ return -1;
+
+ max_w = n * (n + 1) / 2;
+ if (w < max_w / 2)
+ w = max_w - w;
+
+ return count_sums_to_W (n, ceil (w)) / (double) (1 << n) * 2;
+}