Yahya Nur Ifriza, Daffa Ra'if Ridhodin, Subhan Subhan, Devi Ajeng Efrilianda
Breast cancer is one of the most common diseases that affect women and needs to be correctly diagnosed as soon as possible. Using machine learning in medical decision support systems is one potential method to diagnose this illness more accurately. This study sought to address the limitations of classification techniques like Random Forest through the selection of pertinent characteristics. This study focuses on improving feature selection by utilizing the Ant Colony Optimization (ACO) method, aiming to boost the accuracy of breast cancer detection. This work also advances the development of machine learning-based medical decision support systems and addresses the need for a more precise and effective classification model. The dataset utilized was the Wisconsin Diagnostic Breast Cancer (WDBC) from the UCI Machine Learning Repository. Relevant traits were identified through ACO prior to categorization with Random Forest algorithms. Evaluation is based on classification accuracy. The Random Forest algorithm's classification accuracy increased from 96.49% to 99.12% after feature selection using ACO. ACO was also able to reduce the number of characteristics and maximize computational performance. This investigation employs Random Forest alongside a metaheuristic optimization method known as Ant Colony Optimization to enhance the model's precision and effectiveness in identifying breast cancer. © 2026 IEEE.
Universitas Negeri Semarang, Department of Computer Science, Indonesia
