Meta-learning and Bayesian Networks to Construct Classifier Ensembles for Gene-Expression Data

• Grant number: 12/22295-0
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• Start date: March 01, 2013
• End date: September 30, 2015
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: André Carlos Ponce de Leon Ferreira de Carvalho
• Grantee: Edwin Rafael Villanueva Talavera
• Host Institution: Instituto de Ciências Matemáticas e de Computação (ICMC). Universidade de São Paulo (USP). São Carlos , SP, Brazil
• Associated scholarship(s): 14/10852-7 - Detection of functional gene-gene interactions from observational gene expression data using classifier ensembles and Metalearning, BE.EP.PD

Abstract

Machine learning has been successfully used in Bioinformatics. The use of gene-expression data to classify biological tissues into known categories has generated high expectations for improving the diagnosis, prognosis and treatment of severe diseases, like cancer. Nevertheless, building an accurate classifier for such purpose has proven to be a challenging task, given the very high dimensionality of gene-expression data and the noisy nature thereof. Among the most promising approaches to deal with this problem are the so-called ensembles of classifiers, which classify the samples based in the combination of several classifiers, in the hope that many experts can lead to a better decision than a single one. However, the construction of an ensemble is a complex problem and the existing techniques for this end often induce suboptimal and poorly interpretable classification systems. In this research project we intend to study how the advantages of meta-learning and Bayesian Networks can be exploited to facilitate the construction of ensembles of classifiers for gene-expression data, approaches that have been successful in a diversity of fields, but little explored in the problem under study.

Scientific publications

(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)

ESPEZUA, SOLEDAD; VILLANUEVA, EDWIN; MACIEL, CARLOS D.; CARVALHO, ANDRE. A Projection Pursuit framework for supervised dimension reduction of high dimensional small sample datasets. Neurocomputing, v. 149, n. B, p. 767-776, FEB 3 2015. (12/22295-0)

ESPEZUA, SOLEDAD; VILLANUEVA, EDWIN; MACIEL, CARLOS D.; CARVALHO, ANDRE. A Projection Pursuit framework for supervised dimension reduction of high dimensional small sample datasets. Neurocomputing, v. 149, p. 10-pg., 2015-02-03. (12/22295-0)