1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
|
#!/usr/bin/perl
# Copyright (c) 2003, 2005 MySQL AB
#
# 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; version 2 of the License.
#
# 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
# compares out-files (as created by Ereport.pl) from dir1/*.out and dir2/*.out
# for each effectiveness column computes the probability of the hypothesis
# "Both files have the same effectiveness"
# sign test is used to verify that test results are statistically
# significant to support the hypothesis. Function is computed on the fly.
# basic formula is \sum_{r=0}^R C_N^r 2^{-N}
# As N can be big, we'll work with logarithms
$log2=log(2);
sub probab {
my $N=shift, $R=shift;
my $r, $sum=0;
for $r (0..$R) {
$sum+=exp(logfac($N)-logfac($r)-logfac($N-$r)-$N*$log2);
}
return $sum;
}
# log(N!)
# for N<20 exact value from the table (below) is taken
# otherwise, Stirling approximation for N! is used
sub logfac {
my $n=shift; die "n=$n<0" if $n<0;
return $logfactab[$n] if $n<=$#logfactab;
return $n*log($n)-$n+log(2*3.14159265358*$n)/2;
}
@logfactab=(
0, 0, 0.693147180559945, 1.79175946922805, 3.17805383034795,
4.78749174278205, 6.57925121201010, 8.52516136106541, 10.6046029027453,
12.8018274800815, 15.1044125730755, 17.5023078458739, 19.9872144956619,
22.5521638531234, 25.1912211827387, 27.8992713838409, 30.6718601060807,
33.5050734501369, 36.3954452080331, 39.3398841871995, 42.3356164607535,
);
############################# main () ###############################
#$p=shift; $m=shift; $p-=$m;
#if($p>$m) {
# print "1 > 2 [+$p-$m]: ", probab($p+$m, $m), "\n";
#} elsif($p<$m) {
# print "1 < 2 [+$p-$m]: ", probab($p+$m, $p), "\n";
#} else {
# print "1 = 2 [+$p-$m]: ", probab($p+$m, $m), "\n";
#}
#exit;
die "Use: $0 dir1 dir2\n" unless @ARGV==2 &&
-d ($dir1=shift) && -d ($dir2=shift);
$_=`cd $dir1; echo *.out`;
s/\.out\b//g;
$total="";
for $file (split) {
open(OUT1,$out1="$dir1/$file.out") || die "Cannot open $out1: $!";
open(OUT2,$out2="$dir2/$file.out") || die "Cannot open $out2: $!";
@p=@m=();
while(!eof(OUT1) || !eof(OUT2)) {
$_=<OUT1>; @l1=split; shift @l1;
$_=<OUT2>; @l2=split; shift @l2;
die "Number of columns differ in line $.\n" unless $#l1 == $#l2;
for (0..$#l1) {
$p[$_]+= $l1[$_] > $l2[$_];
$m[$_]+= $l1[$_] < $l2[$_];
}
}
for (0..$#l1) {
$pp[$_]+=$p[$_]; $mm[$_]+=$m[$_];
$total[$_].=rep($file, ($#l1 ? $_ : undef), $p[$_], $m[$_]);
}
close OUT1;
close OUT2;
}
for (0..$#l1) {
rep($total[$_], ($#l1 ? $_ : undef), $pp[$_], $mm[$_]);
}
sub rep {
my ($test, $n, $p, $m, $c, $r)=@_;
if ($p>$m) { $c=">"; $r="+"; }
elsif($p<$m) { $c="<"; $r="-"; }
else { $c="="; $r="="; }
$n=" $n: " if defined $n;
printf "%-8s $n $dir1 $c $dir2 [+%03d-%03d]: %16.15f\n",
$test, $p, $m, probab($p+$m, ($p>=$m ? $m : $p));
$r;
}
|
