Web-Based Wine Quality Prediction Using Explainable AI

The quality of red wine can be determined based on several physicochemical and sensory properties. Conventional methods of wine evaluation depend heavily on expert tasters, making the whole process subjective and inconsistent. In this paper, we present a web-based wine quality prediction system based on Explainable Artificial Intelligence (XAI), a powerful tool which can provide clarity and aquinting in the wine quality prediction extraction. We train several machine learning models, including the Random Forest, XGBoost, and Neural Networks, and we evaluate using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> Score. We find that XGBoost performs the best in accuracy, and Random Forest has a balance between interpretability and accuracy.To increase interpretability, we use SHapley Additive exPlanations (SHAP) and Local Interpre table Model-agnostic Explanations (LIME) to evaluate the main features impacting wine quality, including alcohol, volatile acidity, and citric acid levels. Unlike traditional black-box AI models, our system explains to winemakers, sommeliers, and consumers why a given wine is rated the way it is. Moreover, we set up the model as a web application for convenient, real-time predictions of wine quality. This research simultaneously intersects AI and the wine industry by providing a data-driven, interpretable approach to assessing wine quality.