An accuracy adaptive breast tumor gene classification method

Abstract

The method on gene classification has been widely studied with the development of gene chip. Machine learning is the best choice to research the issue. But both traditional SVM and ELM cannot fulfill the requirement of high accuracy and short time. Therefore, in this paper, we propose a novel Accuracy Adaptive Extreme Learning Machine (A2-ELM) which can cover the shortage of traditional SVM and ELM in the fact of more dynamic. Firstly, we propose a method of feature selection and overview the property of traditional ELM. Then, an Accuracy of Adaptive ELM (A2-ELM) is developed, which can fulfill the requirement for accurately and rapidly. Finally, we conduct experiments on gene expression data to verify the dynamic and accurate of our proposed accuracy of adaptive ELM in classification gene expression data with experimental settings.