This software provides a set of Java classes specifically designed for multi-label classification. It helps users achieve better classification results than traditional single-label methods. The package is easy to use and can be integrated into existing Java projects.
Built on top of Weka, Mulan requires the specified version of Weka (currently Weka v.3.5.7) as well as Java v1.5 or better. It's excellent that this software is distributed under the GNU GPL license.
Mulan accurately processes input files in ARFF format structured in the following way: input attributes must be followed by |L| attributes, where L is the set of labels, both of which must accept values from the set {0, 1}. A value of 1 for these attributes will indicate that an instance belongs to this label, and a value of 0 will show that it does not belong to this label.
If you encounter any issues while using Mulan, please contact Grigorios Tsoumakas for bug reports, comments, suggestions, or requests for help with the package. All in all, Mulan is a feature-rich and valuable software tool for multilabel classification.
Version 0.4.0: N/A