Comprehensive analysis of classification, labeling, and schemes for digital twins in manufacturing systems, including unsupervised and supervised learning techniques for anomaly detection, predictive maintenance, and quality control