Anemia prediction using gene expression programming (GEP) and explainable artificial intelligence approaches.
Journal:
Computers in biology and medicine
Published Date:
Jul 31, 2025
Abstract
Anemia being a global health disorder, affecting millions of people, especially pregnant women, children, and the elderly. Proper and timely diagnosis must be ensured to prevent its adverse effects, but the traditional diagnostic methods are very time-consuming, costly, and subject to human mistakes. This research investigates the application of Gene Expression Programming (GEP), a proven technique in machine learning (ML), in predicting anemia. A publicly available dataset on Kaggle was utilized, with clinical parameters including hemoglobin and red blood cell indices. The hyperparameters of the GEP model were best optimized, and the accuracy rate was 99.30 %. To increase the interpretability of the model, Explainable AI methods Local Interpretable Model-agnostic Explanations (LIME) and Shapley Additive exPlanations (SHAP) were used. The findings revealed that hemoglobin levels and gender were the most significant features for predicting anemia. The work brings into limelight the usefulness of ML-based diagnostic solutions in medicine with dependable, automatic, and interpretable models for classifying anemia.
