<![CDATA[Abstract Wine quality is a crucial aspect in assessing the consistency of fermented beverage products. This study compares the performance of four classification algorithms Support Vector Machine (SVM), Naive Bayes, K-Nearest Neighbor (KNN), and Decision Tree in predicting wine quality based on chemical parameters. The research includes feature normalization, binary target transformation, and evaluation using 80:20 and 90:10 data-splitting schemes. The results show that SVM achieves the highest performance with an accuracy of 0.8908 under the 80:20 scheme and remains stable in the 90:10 scheme. Naive Bayes yields the highest recall, making it more effective in identifying high-quality wine, while KNN provides competitive accuracy but low recall. Decision Tree produces the low est performance due to its tendency to overfit the dataset. Overall, SVM is recommended as the most optimal algorithm for wine quality prediction.]]>
Copyrights © 2026
