Wine Quality Detection Based on Improved Stacking Ensemble Learning

• Published in: 2023 8th International Conference on Information Systems Engineering (ICISE) pp. 226 - 229
• Main Authors: Zhou, Mengkun, Yu, Wenhui, Jiang, Kun
• Format: Conference Proceeding
• Language: English
• Published: IEEE 23.06.2023
• Subjects: Base learners; Classification algorithms; Ensemble learning; Glucose; Prediction algorithms; Robustness; Stability analysis; Stacking; Stacking algorithm; Support vector machines; Wine quality detection
• ISSN: 2643-7309
• DOI: 10.1109/ICISE60366.2023.00054

Ensuring customer satisfaction and improving business profitability are crucial for high-quality wines. However, traditional sensory analysis-based wine quality testing can be subjective and cost-prohibitive. Machine Learning can be applied in wine quality evaluation. But the limitations in handling various data distributions or capturing typical nonlinear relationships in wine quality data. Thus, this paper proposes an improved Stacking ensemble learning method that overcomes the limitations of ML models and enhances accurate wine quality prediction. Based on the fundamental Stacking algorithm, the algorithm optimizes the base learners to improve the accuracy of prediction results and improves the traditional Stacking ensemble learning algorithm by introducing the raw data and adding the weighted average of the predictions made by the base learners, explaining its different reliability levels. The algorithm is applied to a publicly available dataset of wine quality and shows improved accuracy and robustness compared to individual base learners or other ensemble methods. The final prediction accuracy reached 91.70%.