Abstract

Gastric disease progression is challenging to predict due to the complex nature of endoscopic images. This study addresses the problem by integrating fuzzy logic with machine learning, specifically XGBoost, for predictive modeling. The proposed method preprocesses endoscopic images, extracts features, and applies fuzzy logic for classification, followed by XGBoost for final prediction. Results demonstrate an accuracy of 92.5% and an F1-score of 0.91, outperforming traditional methods. The model offers a robust tool for early detection and monitoring of gastric diseases, enhancing clinical decision-making.