Title: Minimum redundancy maximum relevance with mean-based ranking for biomarker gene selection in autism
Authors: Prema Ramasamy; Premalatha Kandhasamy
Addresses: MCA Department, New Horizon College of Engineering, Outer Ring Road, Near Marathalli, Bellandur Main Road, Bengaluru, Karnataka 560103, India ' Department of Computer Science and Engineering, Faculty of Information and Communication Technology, Bannari Amman Institute of Technology, Erode, India
Abstract: Autism is the most commonly occurring form of autism spectrum disorder. Analysis of gene expression data is important in autism in order to retrieve the required information. Gene expression data generally contain a large number of genes but a small number of samples. The complicated relations among the different genes make analysis more difficult, and removing the irrelevant genes will improve the quality of the results. This paper presents a feature selection method based on minimum redundancy maximum relevance with mean-based ranking (mRMR-µR) to identify the genes associated with autism or potential susceptibility regions in the genome. To assess the performance of the proposed method, it is applied on autism gene expression data. The classifiers support vector machine (SVM), k-nearest-neighbour (kNN) and artificial neural network (ANN) are used to identify the accuracy of selected features. The experimental results show that the proposed feature selection method gives 100% average classification accuracy for the top twenty-five selected genes.
Keywords: autism; gene expression data; feature selection; classification; minimum redundancy maximum relevance.
International Journal of Telemedicine and Clinical Practices, 2019 Vol.3 No.2, pp.95 - 110
Received: 19 May 2017
Accepted: 03 May 2018
Published online: 04 Jun 2019 *