AI-Based Manufacturing Efficiency Classification Using Sensor, Production, and 6G Network Data: A Validation Study of the Thales Group Manufacturing Dataset

This research evaluates AI-based manufacturing efficiency classification using industrial sensor, production, quality-control, and 6G network telemetry data from the Thales Group manufacturing dataset. It compares Logistic Regression, Random Forest, and Gradient Boosting models using chronological train-test splitting, feature ablation, and permutation importance. Although the Random Forest model achieves 100% accuracy, further analysis reveals that the efficiency labels are almost entirely determined by production speed and error rate. The study highlights synthetic data characteristics, target leakage, class imbalance, and the absence of meaningful relationships between network metrics and manufacturing efficiency, emphasizing the importance of data validation and reliable machine learning evaluation.