Wineinformatics: Uncork Napa´s Cabernet Sauvignon by Association Rule Based Classification

Wineinformatics is a developing data science field regarding the application of data mining on professional wine reviews. In this paper, we propose a wine region specific concept to study terror by collecting 1200 different wine reviews from Napa Valley, California and construct the dataset via the Computational Wine Wheel. We apply association rule based classification algorithm to predict the quality of the wines through attributes extracted from wine reviews. The prediction accuracy of our predictions was satisfactory, frequently reaching the 74% - 76% range, while still maintaining above 90% coverage. Compare with previous research, the much higher coverage proves wines from the same region share similar patterns. To the best of our knowledge, this paper is the first article deals with a single region wine dataset, which is also the largest dataset in any wine related data mining research papers.