Attribute Weighting Model for Breast Cancer Prediction with the Harmony Search Algorithm

Breast cancer is a disease that affects many women worldwide. Identifying risk factors is important for prevention and early treatment. Although models such as Gail, Tyrer-Cuzick, and BOADICEA can predict breast cancer risk at five to ten years based on risk factors, the Gail model has been shown to have poor accuracy. Moreover, accurately assessing the influence of risk factors remains a challenge. Hence, accurate models are needed for early detection. In this paper, we used the harmony search algorithm to assign weights to each risk factor value to produce an accurate predictive model. Our model achieved a 96% precision and an 81% accuracy, outperforming Gail’s results, which obtained a 67% precision and a 60% accuracy. Furthermore, the simplicity of our model makes it a valuable tool for both patients and medical professionals.