Breast cancer is one of the most prevalent diseases among women, and early prediction plays a crucial role in improving clinical outcomes. Body Mass Index (BMI) is one of several factors that have shown reliability in contributing to breast cancer risk assessment. In this study, we propose a multi-task classification framework based on the Directed Multi-Task Twin Support Vector Machine (DMTSVM) to predict breast cancer using BMI-related information. To better capture population heterogeneity, BMI is decomposed into two groups, each treated as a separate task within the multi-task learning structure.The model includes two balancing parameters. To determine these parameters effectively, we employ two metaheuristic optimization algorithms, Particle Swarm Optimization (PSO) and Grey Wolf Optimizer (GWO). This integration enables the model to enhanced predictive performance. Experimental evaluations on a breast cancer dataset with BMI-based tasks demonstrate that the optimized DMTSVM framework outperforms the non-optimized baseline. The results confirm the effectiveness of combining multi-task learning with PSO and GWO based parameter tuning for medical prediction problems.

Keywords