MKBoost: A Framework of Multiple Kernel Boosting

Hao Xia, Steven C. H. Hoi

Abstract

Multiple kernel learning (MKL) is a promising family of machine learning algorithms using multiple kernel functions for various challenging data mining tasks. Conventional MKL methods often formulate the problemas an optimization task of learning the optimal combinations of both kernels and classifiers, which usually results in some forms of challenging optimization tasks that are often difficult to be solved. Different from the existing MKL methods, in this paper, we investigate a boosting framework of multiple kernel learning for classification tasks, i.e., we adopt boosting to solve a variant of MKL problem, which avoids solving the complicated optimization tasks. Specifically, we present a novel framework of Multiple Kernel Boosting (MKBoost), which applies the idea of boosting techniques to learn kernel-based classifiers with multiple kernels for classification problems. Based on the proposed framework, we propose several variants of MKBoost algorithms and extensively examine their empirical performance on a number of benchmark datasets in comparisons to various state-of-the-art MKL algorithms on classification tasks. Experimental results show that the proposed method is more effective and efficient than the existing MKL techniques. 

Publication
  • "MKBoost: A Framework of Multiple Kernel Boosting" Hao Xia, Steven C.H. Hoi,
    IEEE Transactions on Knowledge and Data Engineering (TKDE),  25(7), 1574-1586, 2013.
    [ PDF ] (preprint)

  • "MKBoost: A Framework of Multiple Kernel Boosting" Hao Xia, Steven C.H. Hoi
    Proceedings of the Eleventh SIAM International Conference on Data Mining (SDM-2011), 2011.
    [ PDF ]

  • @article{Xia/Hoi/MKBoost2013,
    title={Mkboost: A framework of multiple kernel boosting},
    author={Xia, Hao and Hoi, Steven CH},
    journal={IEEE Transactions on Knowledge and Data Engineering (TKDE) },
    volume={25},
    number={7},
    pages={1574--1586},
    year={2013},
    publisher={IEEE}
    }

 

Source Code

You need to agree to this policy before downloading and using the code and data below.

MATLAB Source Code for MKBoost: >> Click here to download CODE <<

Note: The code was implemented by Mr. Hao Xia. Feel free to send us your comments and suggestion.
 

Instructions for the MATLAB package


Please refer to "demomain.m" for typical usage. You can run the demo directly by typing "demomain" in matlab command window on a 64bit linux machine.

The package needs LIBSVM package, which can be downloaded from the LIBSVM website (http://www.csie.ntu.edu.tw/~cjlin/libsvm/). If you are not using 64bit linux machine, you may need to recompile it in your machine.

The code can only accept binary labels (-1, +1), it can be extended to solve multi-class classification problem easily by adopting some existing techniques, such as one-to-one approach in our work.

 

Datasets in Our Experiments

MATLAB Datasets for MKBoost: >> Click here to download DATA <<

Instructions for Datasets
These are the 16 datasets used in our experiments. We randomly choose them from web machine learning repositories, which can be downloaded from http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/datasets/. They are in matlab ".mat" format. X is the data matrix (N * dim), Y is the label vector (N*1).

Links to Resources and Acknowledgement
 
Back to Steven Hoi's Homepage