Bootstrap a statistic in a grouped sample
Function BSTRAG builds on an earlier FEX submission, BSTRAP - itself an extension of BOOTSTRP (Statistics Toolbox, v. 5-6) - by implementing 'groupwise' operation, wherein rows of x (,y,z,..) are split into several groups (having n1, n2, n3, .. rows), marked by distinct values of grouping variable c, and each resampling iteration randomly selects n1 rows from group 1, n2 rows from group 2, etc.
Syntax of BSTRAG is similar to that of BSTRAP, but differs in requiring an extra input argument, group-assignment vector, which must also be supplied to the function computing (matrix- or structure-valued) bootstrapped statistics.
Function BSTATS inputs an array or a structure array produced by BSTRAP, and outputs means, variances and (2.5%, 97.5%) percentiles of the bootstrap distributions. (Percentiles are calculated using PRCTILE of Statistics Toolbox - replacing it with a user-contributed analog should be straightforward).
See BSTRAG_DEMO and BSTRAP_DEMO for examples.
引用
Dimitri Shvorob (2024). Bootstrap a statistic in a grouped sample (https://www.mathworks.com/matlabcentral/fileexchange/14411-bootstrap-a-statistic-in-a-grouped-sample), MATLAB Central File Exchange. に取得済み.
MATLAB リリースの互換性
プラットフォームの互換性
Windows macOS Linuxカテゴリ
- AI and Statistics > Statistics and Machine Learning Toolbox > Probability Distributions > Resampling Techniques >
タグ
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!bstrag/
バージョン | 公開済み | リリース ノート | |
---|---|---|---|
1.0.0.0 | BSD |