Wine Quality Prediction using ML Techniques and KNIME

Abstract

The Wine quality is important for purchasers as well as the wine industry to produce in good quantity. The normal way of quantifying wine quality is tedious. These days, machine learning models are key tools in replacing human tasks from measuring alcohol quality. While in quality prediction, there are several features, but not all the traits will not be relevant to quality prediction. Classification of wine quality is a complex work as the Flavour is the least aspect of human senses. For wine quality prediction RFC, SVM, Logistic Regression, GDC and Bayesian classifier demonstrates to be better with greater prediction accuracy than other data mining techniques. This prediction can be used in CART, SVM, Random Forest (RF) and Big-Data. The performance of the proposed model achieved the highest classification accuracy (99%) using Random Forest classifier. The paper explores which of the features wine determines the best quality of wine and generate insights into each of these features.