Share
Export Citation
Selection of Informative Genes to Classify Type 2 Diabetes Mellitus using Support Vector Machine
Auliah F.N.
5th International Conference on Computing Engineering and Design Icced 2019
Abstract
Type 2 Diabetes Mellitus is a group of disorders with characteristics such as insulin resistance, impaired insulin secretion, and increased glucose production, which has patients reaching 90% - 95% of the overall population of people with diabetes mellitus. There at most 70% of Indonesians are unaware attacked by diabetes. Therefore, an early detection has an important role and the utilization of microarray technology can cope this problem. One of the challenges for microarray applications is to select an appropriate number of the most significant genes for data analysis. Besides that, it is hard to accomplish a satisfactory classification results by Support Vector Machine (SVM) due to the dimensionality and the over-fitting problems. For this reason, it is desirable to select informative genes firstly in order to improve classification accuracy of SVM classifier. In this study, we use the Information Gain in order to determine informative features of data to get better classification performances and then the SVM is applied to the selected features. Based on the result, the informative genes were selected from 25,770 to 309 genes. SVM can predict sample with accuracy 100% and area under curve 100%. There is a probe which is a gene that has various functions, including regulating neurotransmitter releases, heart rate, insulin secretion, neuronal excitability, epithelial electrolyte transport, smooth muscle contraction, and cell volume.