Customized fuzzy rough quick reduct for attribute selection in microarray data

Abstract

Attribute selection has gained much of attention in recent years in the field of Bioinformatics Machine learning algorithms are used for the diagnosis and treatment of a number of diseases especially cancer which is considered to be one of the deadliest diseases the world over These algorithms aid in diagnosing the disease at an early stage thereby increasing the survival rate and reducing the mortality rate since traditional approach of diagnosis of the disease is time consuming and error prone due to human intervention The use of computer methods and algorithms to process microarray databases gains importance because of the rapid growth of the size of the database in recent years Microarray data consists of small sample of training and testing samples with high dimensionality All the attributes do not contribute to the cause of cancer The performance of the learning algorithm is deteriorated by the memory occupied by the high dimensional data that consists of a number of irrelevant attributes The main objective of this research work is to identify the informative genes and remove the redundant genes and thereby contribute to the increase in classification accuracy and the decrease in the number of attribute genes and computation time The key motivating factor of research in the area of attribute selection in microarray data is the high dimensionality of the datasets that suffer from the problems of curse of dimensionality. newline