line |
stmt |
bran |
cond |
sub |
pod |
time |
code |
1
|
|
|
|
|
|
|
package Text::NumericData::Stat; |
2
|
|
|
|
|
|
|
|
3
|
|
|
|
|
|
|
# TODO: Integrate into Text::NumericData::File already. |
4
|
|
|
|
|
|
|
# Also, return something for single data sets; just set non-computable measures to something invalid. |
5
|
|
|
|
|
|
|
|
6
|
2
|
|
|
2
|
|
1584
|
use Text::NumericData::File; |
|
2
|
|
|
|
|
6
|
|
|
2
|
|
|
|
|
62
|
|
7
|
|
|
|
|
|
|
|
8
|
2
|
|
|
2
|
|
12
|
use strict; |
|
2
|
|
|
|
|
4
|
|
|
2
|
|
|
|
|
997
|
|
9
|
|
|
|
|
|
|
|
10
|
|
|
|
|
|
|
# This is just a placeholder because of a past build system bug. |
11
|
|
|
|
|
|
|
# The one and only version for Text::NumericData is kept in |
12
|
|
|
|
|
|
|
# the Text::NumericData module itself. |
13
|
|
|
|
|
|
|
our $VERSION = '1'; |
14
|
|
|
|
|
|
|
$VERSION = eval $VERSION; |
15
|
|
|
|
|
|
|
|
16
|
|
|
|
|
|
|
# Generate a statistics file out of given input file. |
17
|
|
|
|
|
|
|
# Gives and takes Text::NumericData::File objects. |
18
|
|
|
|
|
|
|
# This could be a method of Text::NumericData::File, but it doesn't have to. |
19
|
|
|
|
|
|
|
sub generate |
20
|
|
|
|
|
|
|
{ |
21
|
1
|
|
|
1
|
0
|
3
|
my $in = shift; |
22
|
1
|
|
|
|
|
4
|
my $out = Text::NumericData::File->new($in->{config}); |
23
|
1
|
|
|
|
|
8
|
my @mean; # arithmetic mean |
24
|
|
|
|
|
|
|
my @error; # standard error, sqrt(mean sq. error / N-1) |
25
|
1
|
|
|
|
|
0
|
my @min; # minimal value |
26
|
1
|
|
|
|
|
0
|
my @max; # maximal value |
27
|
|
|
|
|
|
|
|
28
|
1
|
|
|
|
|
4
|
$out->{title} = 'Statistics'; |
29
|
1
|
50
|
|
|
|
14
|
$out->{title} .= ' of '.$in->{title} if defined $in->{title}; |
30
|
1
|
|
|
|
|
7
|
$out->{titles} = ['column', 'name', 'mean', 'stderr', 'min', 'max']; |
31
|
|
|
|
|
|
|
|
32
|
1
|
|
|
|
|
2
|
my $N = @{$in->{data}}; |
|
1
|
|
|
|
|
3
|
|
33
|
1
|
50
|
|
|
|
3
|
return $out if $N < 1; |
34
|
1
|
|
|
|
|
2
|
my $S = @{$in->{data}[0]}; |
|
1
|
|
|
|
|
25
|
|
35
|
|
|
|
|
|
|
|
36
|
1
|
|
|
|
|
5
|
for my $i (0 .. $S-1) |
37
|
|
|
|
|
|
|
{ |
38
|
2
|
|
|
|
|
23
|
$max[$i] = $min[$i] = $mean[$i] = $in->{data}[0][$i]; |
39
|
|
|
|
|
|
|
} |
40
|
1
|
|
|
|
|
10
|
for(my $j=1; $j<@{$in->{data}}; ++$j) |
|
10
|
|
|
|
|
21
|
|
41
|
|
|
|
|
|
|
{ |
42
|
9
|
|
|
|
|
15
|
my $d = $in->{data}[$j]; |
43
|
9
|
|
|
|
|
17
|
for my $i (0 .. $S-1) |
44
|
|
|
|
|
|
|
{ |
45
|
18
|
|
|
|
|
25
|
$mean[$i] += $d->[$i]; |
46
|
18
|
50
|
|
|
|
35
|
$min[$i] = $d->[$i] if $d->[$i] < $min[$i]; |
47
|
18
|
100
|
|
|
|
36
|
$max[$i] = $d->[$i] if $d->[$i] > $max[$i]; |
48
|
|
|
|
|
|
|
} |
49
|
|
|
|
|
|
|
} |
50
|
1
|
|
|
|
|
10
|
for my $i (0 .. $S-1) |
51
|
|
|
|
|
|
|
{ |
52
|
2
|
|
|
|
|
5
|
$mean[$i] /= $N; |
53
|
|
|
|
|
|
|
} |
54
|
1
|
|
|
|
|
1
|
for my $d (@{$in->{data}}) |
|
1
|
|
|
|
|
2
|
|
55
|
|
|
|
|
|
|
{ |
56
|
10
|
|
|
|
|
15
|
for my $i (0 .. $S-1) |
57
|
|
|
|
|
|
|
{ |
58
|
20
|
|
|
|
|
39
|
$error[$i] += ($d->[$i]-$mean[$i])**2; |
59
|
|
|
|
|
|
|
} |
60
|
|
|
|
|
|
|
} |
61
|
1
|
|
|
|
|
5
|
for my $i (0 .. $S-1) |
62
|
|
|
|
|
|
|
{ |
63
|
2
|
50
|
|
|
|
10
|
$error[$i] = $N > 1 ? sqrt($error[$i]/($N-1)) : 0; |
64
|
|
|
|
|
|
|
} |
65
|
1
|
|
|
|
|
3
|
for my $s (1 .. $S) |
66
|
|
|
|
|
|
|
{ |
67
|
2
|
|
|
|
|
6
|
my $name = $in->{titles}[$s-1]; |
68
|
2
|
50
|
|
|
|
4
|
if(defined $name) |
69
|
|
|
|
|
|
|
{ |
70
|
0
|
|
|
|
|
0
|
$out->filter_text($name); # play safe since there is no quoting in data (yet?) |
71
|
|
|
|
|
|
|
} |
72
|
|
|
|
|
|
|
else |
73
|
|
|
|
|
|
|
{ |
74
|
2
|
|
|
|
|
8
|
$name = "col$s"; |
75
|
|
|
|
|
|
|
} |
76
|
2
|
|
|
|
|
13
|
$out->{data}[$s-1] = [ $s, $name, $mean[$s-1], $error[$s-1], $min[$s-1], $max[$s-1] ]; |
77
|
|
|
|
|
|
|
} |
78
|
1
|
|
|
|
|
4
|
return $out; |
79
|
|
|
|
|
|
|
} |
80
|
|
|
|
|
|
|
|
81
|
|
|
|
|
|
|
1; |
82
|
|
|
|
|
|
|
|
83
|
|
|
|
|
|
|
__END__ |