Microarray gene expression cancer diagnosis using Machine Learning algorithms

In this paper, we use the extreme Learning Machine (ELM) for cancer classification. We propose a two step method. In our two step feature selection method, we first use a gene importance ranking and then, finding the minimum gene subset form the top-ranked genes based on the first step. We tested our two step method in cancer datasets like Lymphoma data set and SRBCT data set. The results in the Lymphoma data set and SRBCT dataset show our two-step methods is able to achieve 100% accuracy with much fewer gene combination than other published results. The results indicate that ELM produces comparable or better classification accuracies with reduced training time and implementation complexity compared to neural networks methods like Back Propagation Networks, SANN and Support Vector Machine methods. ELM also achieves better accuracy for classification of individual categories.