1-Dimensional Convolution Neural Network Classification Technique for Gene Expression Data

TL;DR: Various filter techniques which reduce the high dimensionality of microarray data and different deep learning classification techniques such as 2-Ddimensional Convolution Neural Network (2D- CNN) and 1-Dimensional CNN (1D-CNN) are discussed.

Abstract: In the field of bioinformatics, the development of computational methods has drawn significant interest in predicting clinical outcomes of biological data, which has a large number of features. DNA microarray technology is an approach to monitor the expression levels of sizable genes simultaneously. Microarray gene expression data is more useful for predicting and understanding various diseases such as cancer. Most of the microarray data are believed to be high dimensional, redundant, and noisy. In recent years, deep learning has become a research topic in the field of Machine Learning (ML) that achieves remarkable results in learning high-level latent features within identical samples. This chapter discusses various filter techniques which reduce the high dimensionality of microarray data and different deep learning classification techniques such as 2-Dimensional Convolution Neural Network (2D- CNN) and 1-Dimensional CNN (1D-CNN). The proposed method used the fisher criterion and 1D-CNN techniques for microarray cancer samples prediction.