Concept drift datasets edited to work with Scikit-multiflow directly.
Real:
Dataset | # Features | # Classes | # Samples |
---|---|---|---|
LUdata | 31 | 2 | 1.901 |
Airlines | 7 | 2 | 539.383 |
Chessweka | 8 | 2 | 503 |
Covtype | 54 | 7 | 581.012 |
Elec | 6 | 2 | 45.312 |
Outdoor Stream | 21 | 40 | 4.000 |
Phishing | 46 | 2 | 11.055 |
Poker | 10 | 10 | 829.201 |
Rialto | 27 | 10 | 82.250 |
Spam | 499 | 2 | 6.213 |
Weather | 8 | 2 | 18.159 |
Artifical:
Dataset | # Features | # Classes | # Samples |
---|---|---|---|
InterchangingRBF | 2 | 15 | 200.000 |
MixedDrift | 2 | 15 | 600.000 |
MovingRBF | 10 | 5 | 200.000 |
Moving Squares | 2 | 4 | 200.000 |
Rotating Hyperplane | 10 | 2 | 200.000 |
SEA Big | 3 | 2 | 100.000 |
Sea Stream | 3 | 2 | 40.000 |
Transient Chessboard | 2 | 8 | 200.000 |
Datasets are taken from:
https://sites.google.com/site/zliobaite/resources-1
https://archive.ics.uci.edu/ml/datasets.php
https://github.com/gditzler/ConceptDriftResources
https://github.com/tegjyotsingh/ConceptDriftInduction
https://github.com/vlosing/driftDatasets
http://www.liaad.up.pt/kdus/products/datasets-for-concept-drift