Predictive Modeling of Wine Quality using Machine Learning Techniques

This paper compares the performance of three machine learning models in wine quality prediction: K Nearest Neighbors, Gradient Boosting, and Extreme Gradient Boosting, using a publicly available dataset where various wine chemical properties were used as features for the models. In evaluating model performance, metrics such as accuracy, precision, recall, F1-score, and RMSE were computed for each model. The results indicated that, compared to KNN and GB, XGB had better accuracy and predictive power. Isolated in this study is the potential use of advanced boosting techniques to achieve higher prediction accuracy for wine quality assessment. These findings can further help the winemaker or the quality controller within the wine industry make better quality predictions. Future studies could be done by incorporating more features and other machine learning algorithms to increase the accuracy of predictions.