Enhanced Factor Based Whale Optimization Algorithm with Improved Weight Based Long Short-Term Memory for Cancer Subtypes Diagnosis

The categorization of cancer subtypes using data from GEs (gene expressions) has gained popularity in recent years. In recent years, many supervised learning techniques, particularly those based on DLTs (deep learning techniques), have categorized cancer subtypes. DFN Forests (Deep Flexible Neural Forests) which incorporate FNTs (Flexible Neural Trees) were used more for the categorization of cancer subtypes. However, prior proposals have problems with execution speeds and precisions. To address this issue, this work proposes EFWOA (Enhanced Factor based Whale optimization Algorithm) along with IWLSTMs (Improved Weight based Long ShortTerm Memories) for identifying cancer subtypes. The phases of dimensionality reduction, feature selections, and classifications make up the proposed cancer diagnosis system. Datasets on GEs are used as the initial input. IICA (Improved Independent Component Analysis) was subsequently used to minimize the dimensions. Gene selection is carried out using the EFWOA to increase classification accuracy. Finally, IWLSTMs are used to diagnose cancer subtypes where experimental findings of this new system show better performances in terms of accuracies, precisions, recalls, and f-measures.