Sparse Linear Multi-criteria Optimization Classifier for Disease Prediction

With the changes in environment and lifestyle, the threats of diseases faced by humanity are increasingly growing. In this context, early disease prediction has become a critical issue in the medical field. However, accurate disease diagnosis based on clinical symptoms remains a highly challenging task for healthcare professionals. To address this challenge, data mining technology has demonstrated significant application value in the field of disease prediction. Currently, the amount of data generated annually in the medical field is growing exponentially, and these vast amounts of medical data provide an important foundation for precision medicine. By utilizing advanced data mining techniques, researchers can extract valuable disease characteristic patterns from large medical datasets and establish reliable predictive models. This study innovatively proposes a Sparse Linear Multi-criteria Optimization Classifier (SLMCOC) model for disease prediction. Comparative experiments with classical models such as Decision Trees and KNN demonstrate that SLMCOC achieves higher prediction accuracy. Moreover, its inherent sparsity enables the identification of critical features for classification, thereby enhancing the interpretability of the prediction results.