Abstract
In order to overcome the limitation on small size of gene datasets, many meta-classification methods which ensemble classifiers from different datasets have been developed. However, due to discrepancies of the characteristics among multiple heterogeneous datasets, the number of common and significant genes is usually small. Instead of matching common genes between heterogeneous datasets, we propose a novel solution, alternative feature mapping approach (AFM), to utilize related and discriminative gene expressions while not necessarily having exact matches. Genes in the training dataset are clustered and mapped to the test dataset as gene groups. Through analyzing the correlation within gene groups, significant genes can be matched and dataset dissimilarity factors can be used as weights for meta-classification. We conducted experiments consisting of 10 heterogeneous datasets with different cancer types and platforms. Our experiments show that classification performance is greatly improved using suitable significant genes selected by AFM, and weight voting method based on AFM provides more reliability for meta-classification.
Original language | English |
---|---|
Title of host publication | WMSCI 2010 - The 14th World Multi-Conference on Systemics, Cybernetics and Informatics, Proceedings |
Pages | 341-346 |
Number of pages | 6 |
Volume | 2 |
Publication status | Published - 1 Dec 2010 |
Event | 14th World Multi-Conference on Systemics, Cybernetics and Informatics, WMSCI 2010 - Orlando, FL, United States Duration: 29 Jun 2010 → 2 Jul 2010 |
Conference
Conference | 14th World Multi-Conference on Systemics, Cybernetics and Informatics, WMSCI 2010 |
---|---|
Country/Territory | United States |
City | Orlando, FL |
Period | 29/06/10 → 2/07/10 |
Keywords
- AFM
- Gene Expression Data
- Heterogeneous and Feature Selection
- Meta-classification
ASJC Scopus subject areas
- Artificial Intelligence
- Information Systems