Applying K-Nearest Neighbors Algorithm for Wine Prediction and Classification
Abstract views: 12 , PDF downloads: 8Abstract
This study evaluates the performance of a machine learning classification model using a confusion matrix to analyze predictions across three distinct classes. The results show the model achieving a high accuracy of 94.44%, indicating reliable classification performance. The confusion matrix highlights that most instances were classified correctly, with minimal misclassifications observed, particularly in Class 1, where some overlap with other classes was evident. The findings suggest that the model effectively distinguishes between well-separated classes while facing minor challenges with overlapping data distributions. To address these issues, potential improvements such as feature engineering, class balancing, and advanced optimization techniques are recommended. The study underscores the importance of confusion matrix analysis as a diagnostic tool for understanding classification errors and guiding model refinement. Additionally, this research emphasizes the role of high-quality datasets, proper model selection, and hyperparameter tuning in achieving optimal classification accuracy. The outcomes provide a basis for further enhancement of machine learning models in applications requiring multi-class classification. By reducing errors and improving model robustness, this approach can contribute to more accurate and reliable decision-making processes across various domains, including healthcare, finance, and natural language processing.
Downloads
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
References
[2] Y. Zhang et al., "A Hybrid Machine Learning Approach for Automated Feature Selection and Classification in High-Dimensional Datasets," IEEE Access, vol. 9, pp. 129375–129388, Oct. 2021.
[3] S. Wang, D. Li, Y. Liu, and X. Yu, "Evaluation of Classification Models Based on Confusion Matrix," Proceedings of the 2022 International Conference on Machine Learning and Cybernetics (ICMLC), Guangzhou, China, Aug. 2022, pp. 154–159.
[4] M. Tan and Q. V. Le, "EfficientNetV2: Smaller Models and Faster Training," in Proceedings of the 38th International Conference on Machine Learning (ICML), 2021, pp. 1231–1250.
[5] A. Abdar et al., "A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges," Information Fusion, vol. 76, pp. 243–297, Apr. 2021.
[6] J. He et al., "AutoML: A Survey of the State-of-the-Art," Knowledge-Based Systems, vol. 229, p. 107347, Nov. 2021.
[7] C. Xu, X. Li, J. Zhao, and Y. Wang, "SMOTE-GAN: An Improved Deep Generative Model for Imbalanced Dataset Learning," IEEE Transactions on Industrial Informatics, vol. 17, no. 11, pp. 7764–7773, Nov. 2021.
[8] K. A. Hamid and H. R. Khosravi, "A Novel Ensemble Model Based on Explainable AI for Classification Problems," IEEE Access, vol. 10, pp. 12345–12359, Mar. 2023.
[9] Z. Zhang, X. Zhou, and J. Li, "Learning Class-Weighted Loss for Classification in Imbalanced Data," Pattern Recognition Letters, vol. 156, pp. 47–55, Jan. 2023.
[10] M. Li et al., "Gradient-Based Feature Selection and Optimization for Enhancing Model Classification Accuracy," in Proceedings of the 2023 IEEE International Conference on Data Science and Advanced Analytics (DSAA), Singapore, Sept. 2023, pp. 89–97.
