Experimental Analysis of Deep Learning Techniques for Leukemia Cancer Classification using Micro Gene Array Dataset

Leukemia is a collection of fast-growing blood cancer diseases which are also heterogeneous in terms of genetic and molecular contribution towards disease etiology resulting in difficulty in diagnosis and treatment. This study explores the application of deep learning techniques to classify leukemia subtypes based on gene expression profiles. Utilizing datasets sourced from the Gene Expression Omnibus database, this research implements and evaluates three deep learning models: Long ShortTerm Memory, Bidirectional Long Short-Term Memory, and a novel architecture termed Extended Long Short-Term Memory), which incorporates custom layers and residual connections for enhanced performance. The Extended Long Short-Term Memory model demonstrated superior performance, achieving a final test accuracy of 98.18 %, outperforming both Long Short-Term Memory and Bi-directional Long Short-Term Memory models. The findings highlight the potential of the Extended Long ShortTerm Memory architecture in improving leukemia classification, suggesting its applicability in clinical diagnostics.