A Deep Learning based Feature Selection Method with Multi Level Feature Identification and Extraction using Convolutional Neural Network

: Increasing popularity of feature selection in bioinformatics has led to the development of novel algorithms using neural networks. The objectives of the adaptation of neural networks architectures are proposed on efficient and optimal model for feature classification and selection. A competitive end unique approach in feature selection is adopted here using a convolutional neural network (CNN). Deep learning approach on feature selection is the novel idea which can contribute to the evolvement of identification process, diagnostic methods etc. The experimental work has given good result of ranking the attributes by building a CNN model. The traditional concept of CNN for classification has transformed to a modern approach for feature selection. Handling millions of data with multiple class identities can only be classified with a multi layers network. The CNN models are trained in completely supervised way with a batch gradient back propagation. The parameters are tuned and optimized to get better build type.