Channel Enhanced Deep Convolution Neural Network based Cancer Classification

Himanshu Sharma and Rohit Agarwal

Cancer is a heterogeneous disease and it consists of various types. In cancer research, it is necessary to diagnosis and prognosis the cancer at early stage. The subsequent clinical management of patients are facilitated by this. The expression level of thousands of genes can be measured simultaneously by the advancements in the microarray technology in a single experiment. Swarm intelligence namely Spider Monkey Optimization algorithm is used to reduce the feature in the newly proposed method. And classification is done using SVM. This model cannot produce accurate classification. A novel method is presented in this work to overcome this issue. The noises in the data are removed by the pre-processing using a Robust Multichip Average method. Modified spider Monkey Optimization algorithm is used for selection of feature and classification is done by the Enhanced Deep convolution neural network.

Volume 11 | 10-Special Issue

Pages: 610-617

DOI: 10.5373/JARDCS/V11SP10/20192849