Optimizations Based Feature Selection Method for Disease Survival Prediction

Abstract

In the realm of survival prediction, identifying relevant features plays a pivotal role in enhancing model accuracy and interpretability. This research proposes a novel feature selection method that leverages the synergies between the Whale Optimization Algorithm (WOA) and Genetic Algorithm (GA) to optimize the selection process. The WOA, inspired by the social behavior of humpback whales, is employed to explore the solution space efficiently, while the GA, inspired by the process of natural selection, is used for refining and evolving potential feature subsets. The proposed hybrid algorithm, termed WOA-GA, introduces a dynamic framework that adaptively adjusts the exploration-exploitation trade-off during the search process. The WOA's exploration capabilities are harnessed in the early stages to efficiently traverse the solution space, while the GA's exploitation capabilities are employed later to fine-tune and evolve promising feature subsets. The synergistic combination of these two optimization techniques aims to mitigate the limitations of each individual algorithm and capitalize on their complementary strengths.