Prediction of Red Wine Quality Using One-dimensional Convolutional Neural Networks
As an alcoholic beverage, wine has remained prevalent for thousands of years, and the quality assessment of wines has been significant in wine production and trade. Scholars have proposed various deep learning and machine learning algorithms for wine quality prediction, such as Support vector machine (SVM), Random Forest (RF), K-nearest neighbors (KNN), Deep neural network (DNN), and Logistic regression (LR). However, these methods ignore the inner relationship between the physical and chemical properties of the wine, for example, the correlations between pH values, fixed acidity, citric acid, and so on. To fill the gap, this paper conducts the Pearson correlation analysis, PCA analysis, and Shapiro-Wilk test on those properties and incorporates 1D-CNN architecture to capture the correlations among neighboring features. In addition, it implemented dropout and batch normalization techniques to improve the robustness of the proposed model. Massive experiments have shown that our method can outperform baseline approaches in wine quality prediction. Moreover, ablation experiments also demonstrate the effectiveness of incorporating the 1-D CNN module, Dropout, and normalization techniques.