Swarm intelligence-based neural network applied in predicting breast cancer

Abstract

This paper investigates the effectiveness of a swarm intelligence algorithm namely the time-varying binary particle swarm algorithm in finding an optimal subset of the breast cancer dataset’s features. After the feature selection phase, an artificial neural network was used for predicting the presence of malignant lesions in female breasts. Empirical results show that our approach achieved significant performance in the breast cancer prediction task: Accuracy of 0.9807 (±0.01), Precision of 0.9766 (±0.02), Recall of 0.9728 (±0.02), and F1-score of 0.9742 (±0.01).