A Multi-Objective Feature Selection Framework for Breast Cancer Data Using Evolutionary Methods

Feature selection is a critical component in the development of predictive models for medical data, where high-dimensionality and the need for interpretability pose significant challenges. This research proposes a novel multi-objective feature selection framework that integrates domain expert knowledge with data-driven insights from machine learning models, specifically using random forest feature selection. By assigning weights to features based on expert ratings and random forest importance scores, we employ our proposed algorithm to optimize feature subsets. The proposed method addresses both the dimensionality reduction and accuracy improvement objectives, ensuring the selected features are not only predictive but also clinically relevant. Empirical validation on various medical datasets demonstrates the effectiveness of this approach in enhancing classification performance while maintaining model interpretability. This research bridges the gap between domain expertise and computational feature selection, offering a robust solution for complex medical data analysis.