An Integrated Approach Based on Fuzzy Logic and Machine Learning Techniques for Reliable Wine Quality Prediction

Abstract

In recent decades, wine quality has been one of the predominant problems in many wine industries. However, the analysis of wine quality is inherently complex owing to its multivariate characteristics and the instigation of several sensory features. Most of the existing prediction methods lagged in providing higher detection accuracy for multi-dimensional datasets. To overcome this, a novel Adaptive Wine Quality Prediction (AWQP) approach has been proposed to assess the quality of the wine in an accurate manner. The proposed AWQP methodology entails the development of a Hybrid detection model that encompasses the fuzzy logic principles with the predictive influence of machine learning techniques. Primarily, the sensory features like aroma, taste, and color are delineated by exemplifying the linguistic variables. The fuzzy rules are then determined to collare the qualitative relationships among these different variables. Subsequently, the finest machine learning algorithm can be carried out to train and test the prediction model. The proposed AWQP methodology ameliorates the comprehensibility of the decision-making process through the hybridization of fuzzy logic and the finest machine learning. This proper hybridization facilitates the proposed method to achieve a superior detection accuracy of 98.75% when compared to the existing prediction methods.