Certain investigations on feature Selection techniques for leukaemia Prediction using microarray gene data

Abstract

Cancer identification and classification are the major criteria to be newlineconcerned in the biomedical research. Due to the evolution of advancement in newlinegenetic expressions, the identification of genes, that are vulnerable towards newlinethe cancer occurrence, has become more feasible. Microarray technology, newlinewhich deals with each and every individual gene, has a major impact on newlinecancer treatment. The curse of dimensionality is the main focus to be newlineconsidered while dealing with microarray technique. newlineMicroarray technology deals with vast amount of genetic data. Data newlineanalytics is a major area which deals with large amount of data to identify the newlineinteresting interrelationships within the data. It also identifies the relevant newlinepatterns within the data to obtain valuable information from the data. newlineClassification is an important aspect of data analytics and it deals with both newlinelabelled and unlabelled data. Before performing the classification of data, the newlinerelevant features have to be extracted so that, the error rate during newlineclassification can be minimized. The data analytics techniques are very much newlineuseful to extract valuable resources from genetic expressions. newlineExisting methodologies quoted in literature have performed various newlineoptimization techniques to extract information from genetic expressions and newlinethey are useful in leukaemia predictions. The proposed investigation utilizes newlineAnt Lion optimization and Ant colony optimization as a base and a novel newlinehybridization of ACO, ALO and Particle swarm optimization has been newlineperformed to extract the valuable features. newline