Enhancing Cancer Prediction: Investigating ACO-Based Feature Selection with Logistic Regression, XGBoost, and LightGBM

This report mainly concentrates on cancer prediction, particularly focusing on breast cancer prediction, which remains a significant challenge in healthcare. Improving patient outcomes requires early identification and machine learning approaches have enough opportunities in this area to improve the predicting capabilities. Feature selection plays a vital role in developing accurate predictive models, and this study investigates the effectiveness of Ant Colony Optimization (ACO) in enhancing feature selection for three commonly used machine learning algorithms: Logistic Regression, XGBoost, and LightGBM. The report begins with an introduction highlighting the importance of early cancer detection and the role of machine learning techniques in achieving the goal. Then presents a comprehensive literature review, summarizing previous research papers on cancer detection and classification using various machine learning approaches, with and without implementing ACO. The proposed methodology outlines the steps followed in the research process, from importing necessary libraries to evaluating the performance of machine learning models.