CFS with Combined Search Methods for Dimensionality Reduction in Classifying Aging and Not-Aging related DNA Repair Genes

摘要

Aging is an imperative process, which plays a key role in the development of living organisms. Aging characterizes the growth, lifespan and other important biological phenomena of an organism. Though, aging is common in all organisms, its expression and effect changes with respect to every other organism, which signifies that aging is a complex process associated with many features. DNA damage plays a crucial role in aging and demands the need for classifying DNA repair genes into aging and not-aging related. The amount of information required in classification problem is very huge, leading to the curse of dimensionality. The classification models built will be difficult to understand and analyze because of the dimensionality problem. In this paper we have employed Correlation based Feature subset selection (CFS) along with several search methods to address the curse of dimensionality. The CFS along with combined search methods is aimed in achieving dimensionality reduction by considering only relevant features for classification task. The accuracy of the classification models is considerably better when compared with original data set.