Multi-view multi-class classification for identification of pathogenic bacterial strains
conference paper
In various learning problems data can be available in different representations, often referred to as views. We propose multi-class classification method that is particularly suitable for multi-view learning setting. The algorithm uses co-regularization and error-correcting techniques to leverage information from multiple views and in our empirical evaluation notably outperforms several state-of-the-art classification methods on publicly available datasets. Furthermore,we apply the proposed algorithm for identification of the pathogenic bacterial strains from the recently collected biomedical dataset. Our algorithm gives a low classification error rate of 5%, allows rapid identification of the pathogenic microorganisms, and can aid effective response to an infectious disease outbreak. © Springer-Verlag 2013.
TNO Identifier
488283
ISSN
03029743
ISBN
9783642380662
Source title
11th International Workshop on Multiple Classifier Systems, MCS 2013, 15 May 2013 through 17 May 2013, Nanjing
Pages
61-72
Files
To receive the publication files, please send an e-mail request to TNO Repository.