A hybrid gene selection algorithm for microarray cancer classification using genetic algorithm and learning automata. (2017)

• Record Type: Journal Article
• Authors: Motieghader, Habib; Najafi, Ali; Sadeghi, Balal; Masoudi-Nejad, Ali
• Is Part Of: Informatics in medicine unlocked. Volume 9(2017)
• Page Start: 246
• Page End: 254
• Publication Date: 2017
• Subjects: Biomarker -- Cancer classification -- Gene selection -- Genetic algorithm -- Learning automata; Medical informatics -- Periodicals; 610.285
• DOI: 10.1016/j.imu.2017.10.004
• Languages: English
• ISSNs: 2352-9148

Abstract

Abstract: Cancer classification is an important problem in cancer diagnosis and treatment. One of the most effective methods in cancer classification is gene selection. However, selecting a subset of genes which increases the classification accuracy is an NP-Hard problem. A variety of algorithms were proposed for gene selection in cancer classification in previous studies. In this study, a hybrid meta-heuristic algorithm, which is an integration of Genetic Algorithm and Learning Automata (GALA), is proposed for this purpose. The time complexity of GALA is O ( G . m . n 3 ) and it has acceptable accuracy and performance on some well-known cancer datasets. To evaluate the performance of GALA, six different cancer datasets including Colon, ALL_AML, SRBCT, MLL, Tumors_9 and Tumors_11 were selected. Based on the evaluation process, the GALA algorithm provided remarkable results on each dataset compared to some recently proposed algorithms. Highlights: Cancer classification is an important problem in cancer diagnosis and treatment. In this paper, a hybrid meta-heuristic algorithm, called GALA, is proposed for cancer classification. GALA algorithm uses both genetic algorithm and learning automata advantages. Six cancer microarray datasets (Colon, ALL_AML, SRBCT, MLL, Tumors_9 and Tumors_11) used for evaluation of GALA algorithm.