/BinaryRelevance-Based

The implementation of the paper 'Classifier chains for multi-label classification' in ML 2011 and 'Bayes optimal multilabel classification via probabilistic classifier chains' in ICML 2010. BR-based multi-label algorithm(BR,CC,ECC,PCC)

Primary LanguagePython

BinaryRelevance Based

Dataset

http://mulan.sourceforge.net/datasets-mlc.html

yeast

name domain instances nominal numeric labels cardinality density distinct
yeast biology 2417 0 103 14 4.237 0.303 198

Evaluation

evaluation criterion BR CC ECC PCC(效果很差)
hamming loss 0.2268266085059978 0.2268266085059978 0.23298021498675806 0.5221218258295685
ranking loss 0.16849462724050177 0.16860606695590194 0.045019261908574866
one error 0.24532453245324531 0.25192519251925194 0.24972737186477645

Requrements

  • Python 3.6
  • numpy 1.13.3
  • scikit-learn 0.19.1

Parameter

  • ECC algorithm chain number:10
  • ECC algorithm subset proportion:0.75

Reference

Jesse Read·Bernhard Pfahringer·Geoff Holmes·Eibe Frank, “Classifier chains for multi-label classification,” Machine Learning, vol. 85, no. 3, pp. 333–359, 2011

K. Dembczy´nski, W. Cheng, and E. H¨ullermeier, “Bayes optimal multilabel classification via probabilistic classifier chains,” in Proceedings of the 27th International Conference on Machine Learning, Haifa, Israel, 2010, pp. 279–286