Date of Award

Summer 2005

Document Type

Thesis

Degree Name

Master of Science in Computer Science - (M.S.)

Department

Computer Science

First Advisor

Qun Ma

Second Advisor

Frank Y. Shih

Third Advisor

Barry Cohen

Abstract

In recent years feedforward artificial neural networks (ANN) and their training algorithms have become an effective methodology for the construction of nonlinear systems that solve the statistical problem of classification. The ability of ANNs to solve this problem is highly germane to making progress in the refinement of DNA microarray analysis and techniques regarding this issue. This study attempts to deal with the classification of microarray data and the comparison and validation of simple feedforward ANNs in partitioning high dimensional data. In doing this the efficacy of using ANNs as a genotyping tool will be proven. Furthermore, it has been determined through extensive testing that the classification abilities of simple feedforward ANNs are at least comparable with that of SVMs.

Share

COinS