Training of ANFIS Network by Genetic Algorithm for Diagnosis of Leukemia Cancer Subtypes Using Gene Expression Profile

2017-05-12
Arslan, Mustafa Turan
Haznedar, Bülent
Kalınlı, Adem
In this study, subtypes of Leukemia cancer has classified by using microarray gene expression profiles. An approach is proposed to train Adaptive Neuro Fuzzy Inference System (ANFIS) network by using a population-based Genetic Algorithm (GA) to classify this cancer data. The classification success of the proposed model has compared with the successes of Backpropagation (BP)-ANFIS and Hybrid-ANFIS, which are derivative based ANFIS models. According to obtained results, GA-ANFIS model has performed very well on leukemia cancer, with 90.91% in the classification study. For the same data, BPANFIS and Hybrid-ANFIS models have performed poorly with 63.63% and 59.09%, respectively.
Citation Formats
M. T. Arslan, B. Haznedar, and A. Kalınlı, “Training of ANFIS Network by Genetic Algorithm for Diagnosis of Leukemia Cancer Subtypes Using Gene Expression Profile,” İstanbul, Türkiye, 2017, p. 19, Accessed: 00, 2021. [Online]. Available: https://hdl.handle.net/11511/73791.