Gene selection is an eminent area of research in computational biology. Identifying disease critical genes analysing large gene expression dataset is a challenging problem in computational biology. Application of meta-heuristics is considered to be efficient to attempt such problem. Particle Swarm Optimization (PSO) is a meta-heuristic technique based on swarm intelligence. An improved version of PSO, namely, New Variant Repository and Mutation based PSO (NVRMPSO) combined with kNN (NVRMPSO-kNN) has been applied here to find the subset of genes (disease critical) from gene expression dataset. The modification has been done on a recent PSO version, referred as, Repository and Mutation based PSO (RMPSO). The kNN algorithm has been used for sample classification. The performance of proposed NVRMPSO-kNN has been compared with RMPSO incorporated with kNN (RMPSO-kNN) in gene selection problem. The RMPSO-kNN and proposed NVRMPSO-kNN have been applied to three different sizes of datasets, numbered as, E-MEXP-1050, GSE60438 and GSE10588 of preeclampsia disease. The experimental results reveal that the efficacy of NVRMPSO-kNN is better than that of RMPSO-kNN when Classification Accuracy is considered. Furthermore, the performance of proposed NVRMPSO-kNN has been validated with the state-of-the-art observations.
CITATION STYLE
Jana, B., & Acharyaa, S. (2019). Critical Gene Selection by a Modified Particle Swarm Optimization Approach. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11942 LNCS, pp. 165–175). Springer. https://doi.org/10.1007/978-3-030-34872-4_19
Mendeley helps you to discover research relevant for your work.