Uncertainty Quantification of Multi-Source Information Fusion Model

In modern industrial systems, the accuracy and reliability of fault diagnosis are crucial for ensuring the safe and efficient operation of equipment. Although information fusion technology can enhance the accuracy of fault classification, evaluating diagnostic effectiveness based solely on classification accuracy is insufficient. To ensure reliable estimates of uncertainty that indicate the trustworthiness of model predictions, it is also crucial to dynamically evaluate the data quality from various information sources. To address these issues, this article presents a trustworthy association-based multi-source information fusion classification and uncertainty quantification algorithm. This approach enables reliable classification decisions on multi-source information. Experimental results on multiple public datasets and complex fault diagnosis task demonstrate that this method offers significant advantages in accuracy, robustness, and cred-ibility compared to various advanced methods, confirming its effectiveness in practical fault diagnosis application.