:The package includes the MATLAB code of algorithm MLCT(work on 2 views), which is mainly designed to deal with two issues in multi-label co-training:
1.the widely-witndessed class-imbalance problem of multi-label learning.
2.select the samples and communicate their predicted labels with confidence among multiple co-training classifiers.
This package is free for academic usage. You can run it at your own disk.
This package was developed by Yu-Ying Xing(firstname.lastname@example.org), for any problem concerning the code, please feel free to contact Xing.
For other purposes, please contact Prof. Guo-Xian Yu(email@example.com).
The multi-view multi-label datasets used in this paper can be available at:http://lear.inrialpes.fr/people/guillaumin/data.php
:The package was developed with MATLAB.