time-to-botec

Benchmark sampling in different programming languages
Log | Files | Refs | README

README.md (3383B)


      1 <!--
      2 
      3 @license Apache-2.0
      4 
      5 Copyright (c) 2020 The Stdlib Authors.
      6 
      7 Licensed under the Apache License, Version 2.0 (the "License");
      8 you may not use this file except in compliance with the License.
      9 You may obtain a copy of the License at
     10 
     11    http://www.apache.org/licenses/LICENSE-2.0
     12 
     13 Unless required by applicable law or agreed to in writing, software
     14 distributed under the License is distributed on an "AS IS" BASIS,
     15 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
     16 See the License for the specific language governing permissions and
     17 limitations under the License.
     18 
     19 -->
     20 
     21 # padjust
     22 
     23 > Adjust supplied p-values for multiple comparisons.
     24 
     25 <section class="usage">
     26 
     27 ## Usage
     28 
     29 ```javascript
     30 var padjust = require( '@stdlib/stats/padjust' );
     31 ```
     32 
     33 #### padjust( pvals, method\[, comparisons] )
     34 
     35 Adjusts supplied p-values for multiple comparisons via a specified method.
     36 
     37 ```javascript
     38 var out = padjust( [ 0.1496, 0.0275, 0.3053, 0.1599, 0.2061 ], 'bonferroni' );
     39 // returns [ 0.748, ~0.138, ..., ~0.799, 1 ]
     40 ```
     41 
     42 The `method` parameter can be one of the following values:
     43 
     44 -   **bh**: Benjamini-Hochberg procedure controlling the False Discovery Rate (FDR).
     45 -   **bonferroni**: Bonferroni correction fixing the family-wise error rate by multiplying the p-values with the number of comparisons. The Bonferroni correction is usually a too conservative adjustment compared to the others.
     46 -   **by**: Procedure by Benjamini & Yekutieli for controlling the False Discovery Rate (FDR) under dependence.
     47 -   **holm**: Hommel's method controlling family-wise error rate. It is uniformly more powerful than the Bonferroni correction.
     48 -   **hommel**: Hommel's method, which is valid when hypothesis tests are independent. It is more expensive to compute than the other methods.
     49 
     50 ```javascript
     51 var pvalues = [ 0.319, 0.201, 0.4, 0.374, 0.113 ];
     52 var out = padjust( pvalues, 'holm' );
     53 // returns [ ~0.957, 0.804, ..., ~0.957, ~0.565 ]
     54 
     55 out = padjust( pvalues, 'bh' );
     56 // returns [ 0.4, 0.4, ..., 0.4, 0.4 ]
     57 ```
     58 
     59 By default, the number of comparisons for which the p-values should be
     60 corrected is equal to the number of provided p-values. Alternatively, it is
     61 possible to set `comparisons` to a number greater than the length of
     62 `pvals`. In that case, the methods assume `comparisons - pvals.length`
     63 unobserved p-values that are greater than all observed p-values (for Holm's
     64 method and the Bonferroni correction) or equal to `1` for the remaining methods.
     65 
     66 ```javascript
     67 var pvalues = [ 0.319, 0.201, 0.4, 0.374, 0.113 ];
     68 var out = padjust( pvalues, 'bh', 10 );
     69 // returns [ 0.8, 0.8, ..., 0.8, 0.8 ]
     70 ```
     71 
     72 </section>
     73 
     74 <!-- /.usage -->
     75 
     76 <section class="examples">
     77 
     78 ## Examples
     79 
     80 <!-- eslint no-undef: "error" -->
     81 
     82 ```javascript
     83 var padjust = require( '@stdlib/stats/padjust' );
     84 
     85 var pvalues = [ 0.008, 0.03, 0.123, 0.6, 0.2 ];
     86 
     87 var out = padjust( pvalues, 'bh' );
     88 // returns [ 0.04, 0.075, ~0.205, 0.6, 0.25 ]
     89 
     90 out = padjust( pvalues, 'bonferroni' );
     91 // returns [ 0.04, 0.15, 0.615, 1.0, 1.0 ]
     92 
     93 out = padjust( pvalues, 'by' );
     94 // returns [ ~0.457, ~0.856, 1.0, 1.0, 1.0 ]
     95 
     96 out = padjust( pvalues, 'holm' );
     97 // returns [ 0.2, 0.6, 1.0, 1.0, 1.0 ]
     98 
     99 out = padjust( pvalues, 'hommel' );
    100 // returns [ 0.16, 0.6, 1.0, 1.0, 1.0 ]
    101 ```
    102 
    103 </section>
    104 
    105 <!-- /.examples -->
    106 
    107 <section class="references">
    108 
    109 </section>
    110 
    111 <!-- /.references -->
    112 
    113 <section class="links">
    114 
    115 </section>
    116 
    117 <!-- /.links -->