RankData

Version 1.0.1 (2,67 KB) von thosman
Matlab function to rank each element of data.
39 Downloads
Aktualisiert 13. Mär 2022

RankData

rank = RankData(data, alongDim, invalidRank)

Matlab function to rank each element of data.

Handles all data types accepted by matlab function unique.

Inspiration

Inspired by procrastination and this pick-of-the-week blog post to create a simple function that ranks data but simpler, much faster, and without the limitations from the previous submission.

Params

  • alongDim [default []] - If empty, rank across all entries. If not empty, rank along entries for the specified dim
  • invalidRank [default NaN] - Value invalid entries are set to.

Examples

Example 1

data = randn(2,3,4);
rank = RankData( data )

Example 2

Shows how repeated and invalid entries are handled.

data = [10 nan 11 nan nan 1 2 5 1 7];
rank = RankData(data)

Example 3

Shows how data types like categorical and its invalid data can be naturally handled.

data = randn(5);
data = discretize(data,linspace(0,1,10),'categorical');
rank = RankData(data)

Example 4

Output similar to tiedrank. While we demonstrate matched results, RankData can handle non numeric entries.

data = rand(3,4,2);
data(3,:,:) = data(2,:,:); % Match entries to force ties
alongDim = 1;
rank = tiedrank(RankData(data,alongDim));
isequal(rank,tiedrank(data))

Zitieren als

thosman (2024). RankData (https://github.com/tommyhosman/RankData), GitHub. Abgerufen .

Kompatibilität der MATLAB-Version
Erstellt mit R2019b
Kompatibel mit allen Versionen
Plattform-Kompatibilität
Windows macOS Linux
Tags Tags hinzufügen

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Versionen, die den GitHub-Standardzweig verwenden, können nicht heruntergeladen werden

Version Veröffentlicht Versionshinweise
1.0.1

Added alongDim

1.0.0

Um Probleme in diesem GitHub Add-On anzuzeigen oder zu melden, besuchen Sie das GitHub Repository.
Um Probleme in diesem GitHub Add-On anzuzeigen oder zu melden, besuchen Sie das GitHub Repository.