Research on Anomaly Detection of Parts in Workshop Production Line Based on BO-XGBoostLSS

The management of part quality anomaly events in the workshop production process is an important part of product quality management and one of the main sources of problems in product quality management. To solve the problem of detecting part anomalies in workshop production lines, this paper introduces a new XGBoost framework, the Bayesian optimization-based XGBoostLSS synthesis (BO-XGBoostLSS), which predicts the entire conditional distribution of a univariate response variable, augmenting the existing tree-based gradient enhancement implementations and allows the creation of probabilistic predictions from which prediction intervals and quantiles of interest can be derived, greatly improving the flexibility of XGBoost and the localization accuracy of the test. Experimental results show that the Bayesian optimization-based XGBoostLSS (BO-XGBoostLSS) model achieves low errors on both the training and test sets while achieving good results in terms of both flexibility and prediction accuracy.