Building a Classification Model based on Feature Engineering for the Prediction of Wine Quality by Employing Supervised Machine Learning and Ensemble Learning Techniques

In today’s world, consumers are more concerned regarding the quality of any product. Different approaches are being deployed by various industries to guarantee the excellent quality of their products. Thus, quality certification serve as a vital authentication mechanism for majority of the industries for promoting their numerous products in the market. In the past, only human specialists were employed to evaluate and measure quality. But, nowadays, the majority of validation jobs are automated by software, which curtails the workload of human experts by assisting them to predict the quality of the product and thereby leading to a considerable amount of time saving. Over the past few decades, there has been a sharp rise in wine consumption due to its intrinsic health benefits, particularly for the human heart, as well as for recreational reasons. The main focus of this study is two-fold: the first objective is to predict the quality of wine based upon the correlation between the various physicochemical factors in order to determine the most prominent factors which play a significant role for determining the quality of wine by implementing several supervised machine learning and ensemble learning techniques and the final results being confirmed by employing a variety of quantitative indicators and the second objective is the classification of wine into 3 categories, namely, Best, Good and Poor in order to rank the quality of wine. However, during testing the models with the test dataset, it has been observed that the Random Forest classifier outperformed the other machine learning classifiers with an accuracy of 98%.