P-value correction: Difference between revisions

 

Link with <code>-lm</code>

#include <stdlib.h>//qsort

#include <math.h>//fabs

return 0;

}

 

{{out}}

{{trans|Kotlin}}

To avoid licensing issues, this version is a translation of the Kotlin entry (Version 2) which is itself a partial translation of the Raku entry. If using gcc, you need to link to the math library (-lm).

#include <stdlib.h>

#include <math.h>

each_i(0, 8) adjusted(p_values, types[i]);

return 0;

 

{{output}}

=={{header|C sharp|C#}}==

{{trans|Java}}

using System.Collections.Generic;

using System.Linq;

}

}

{{out}}

<pre>[ 1] 6.126681E-001 8.521710E-001 1.987205E-001 1.891595E-001 3.217789E-001

=={{header|C++}}==

{{trans|Java}}

#include <functional>

#include <iostream>

 

return 0;

{{out}}

<pre>[ 1] 0.6126681081 0.8521710465 0.1987205200 0.1891595417 0.3217789286

{{trans|Kotlin}}

''This work is based on R source code covered by the '''GPL''' license. It is thus a modified version, also covered by the GPL. See the [https://www.gnu.org/licenses/gpl-faq.html#GPLRequireSourcePostedPublic FAQ about GNU licenses]''.

import std.conv;

import std.math;

writefln("\ntype %d = '%s' has a cumulative error of %g", type, types[type], error);

}

{{out}}

<pre>[ 1] 6.126681e-01 0.8521710465 0.1987205200 0.1891595417 0.3217789286

=={{header|Go}}==

{{trans|Kotlin (Version 2)}}

 

import (

fmt.Println(s)

}

 

{{out}}

{{works with|Java|8}}

''This work is based on R source code covered by the '''GPL''' license. It is thus a modified version, also covered by the GPL. See the [https://www.gnu.org/licenses/gpl-faq.html#GPLRequireSourcePostedPublic FAQ about GNU licenses]''.

import java.util.Comparator;

 

}

}

{{out}}

<pre>[ 1] 6.126681e-01 0.8521710465 0.1987205200 0.1891595417 0.3217789286

 

type 5 = 'hommel' has a cumulative error of 4.35302e-07</pre>

 

=={{header|Julia}}==

 

p = [4.533744e-01, 7.296024e-01, 9.936026e-02, 9.079658e-02, 1.801962e-01,

println("\n", corr)

printpvalues(adjust(p, corr))

 

{{out}}

''This work is based on R source code covered by the '''GPL''' license. It is thus a modified version, also covered by the GPL. See the [https://www.gnu.org/licenses/gpl-faq.html#GPLRequireSourcePostedPublic FAQ about GNU licenses]''.

 

 

import java.util.Arrays

println(f.format(type, types[type], error))

}

 

{{out}}

 

To avoid licensing issues, this version follows the approach of the Raku entry of which it is a partial translation. However, the correction routines themselves have been coded independently, common code factored out into separate functions (analogous to Raku) and (apart from the Šidák method) agree with the Raku results.

 

typealias DList = List<Double>

 

types.forEach { println(adjusted(pValues, it)) }

 

{{out}}

=={{header|Nim}}==

{{trans|Kotlin (Version 2)}}

 

type

 

for ctype in CorrectionType:

 

{{out}}

{{trans|C}}

''This work is based on R source code covered by the '''GPL''' license. It is thus a modified version, also covered by the GPL. See the [https://www.gnu.org/licenses/gpl-faq.html#GPLRequireSourcePostedPublic FAQ about GNU licenses]''.

 

use strict;

printf("type $method has cumulative error of %g.\n", $error);

}

 

{{out}}

 

=={{header|Phix}}==

{{out}}

Matches Kotlin (etc) when some of those lines just above are uncommented.

{{trans|Perl}}

''This work is based on R source code covered by the '''GPL''' license. It is thus a modified version, also covered by the GPL. See the [https://www.gnu.org/licenses/gpl-faq.html#GPLRequireSourcePostedPublic FAQ about GNU licenses]''.

import sys

 

error += abs(q[i] - correct_answers[key][i])

print '%s error = %g' % (key.upper(), error)

 

{{out}}

The '''p.adjust''' function is built-in, see [https://stat.ethz.ch/R-manual/R-devel/library/stats/html/p.adjust.html R manual].

 

8.752257e-01, 2.922222e-01, 9.115421e-01, 4.355806e-01, 5.324867e-01,

4.926798e-01, 5.802978e-01, 3.485442e-01, 7.883130e-01, 2.729308e-01,

 

p.adjust(p, method = 'hommel')

 

{{out}}

{{works with|Rakudo|2019.03.1}}

 

 

sub adjusted (@p, $type) { "\n$type\n" ~ format adjust( check(@p), $type ) }

{

say adjusted @p-values, $_

 

{{out}}

=={{header|Ruby}}==

{{trans|Perl}}

x = 1

pmin_array = []

puts "total error for #{method} = #{error}"

end

{{out}}

<pre>Benjamini-Yekutieli

Hommel

total error for Hommel = 1.1483094955369324e-07

</pre>

 

=={{header|SAS}}==

 

input raw_p @@;

cards;

 

proc multtest pdata=pvalues bon sid hom hoc holm;

 

'''output'''

First, install the package with:

 

 

Given a dataset containing the p-values in a variable, the qqvalue command generates another variable with the adjusted p-values. Here is an example showing the result with all implemented methods:

 

 

#delimit ;

}

 

 

'''output'''

50. | .00573549 .2867745 .24993848 .21794862 .19633763 .21794862 .02205958 .09925057 |

+-----------------------------------------------------------------------------------------------+</pre>

 

=={{header|zkl}}==

''This work is based on R source code covered by the '''GPL''' license. It is thus a modified version, also covered by the GPL. See the [https://www.gnu.org/licenses/gpl-faq.html#GPLRequireSourcePostedPublic FAQ about GNU licenses]''.

 

psz,pszf := pvalues.len(), psz.toFloat();

n_i := psz.pump(List,'wrap(n){ pszf/(psz - n) }); # N/(N-0),N/(N-1),..

}

psz.pump(List,'wrap(n){ pa[ro[n]] }); // Hommel q-values

4.533744e-01, 7.296024e-01, 9.936026e-02, 9.079658e-02, 1.801962e-01,

8.752257e-01, 2.922222e-01, 9.115421e-01, 4.355806e-01, 5.324867e-01,

}

println();

{{out}}

<pre style="height:45ex">