[Home]

M³MIML

Description: M3MIML is a package for learning from multi-instance multi-label examples by maximum margin strategy. The package includes the MATLAB code of the algorithm M3MIML, which is designed to deal with multi-instance multi-label learning. It is in particular useful when a real-world object is represented by multiple instances and associated with multiple labels simultaneously.

References:  

• M.-L. Zhang and Z.-H. Zhou. M³MIML: A maximum margin method for multi-instance multi-label learning. In: Proceedings of the 8th IEEE International Conference on Data Mining (ICDM'08), Psia, Italy, 2008.

ATTN:   The main functions of this package is "M3MIML_train.m" and "M3MIML_test.m". You can type "help M3MIML_train" and "help M3MIML_test" in Matlab prompt to find guidance on calling the training and test functions. This package is free for academic usage, you can run it at your own risk.

Requirement: The package was developed with MATLAB. The optimization toolbox needs to be installed to run this package.

ATTN2:       This package was developed by Dr. Min-Ling Zhang (zhangml@hhu.edu.cn). Any problem, please feel free to contact him.

Download:   [code] (7.2Kb)