Real-Time Anomaly Detection of Precision Instruments Based on Robust Random Cut Forest

With the rapid development of the instrument industry, high-precision electrical instruments such as multimeters, digital source meters and nanovoltmeters are increasingly used in research and production, especially in the fields of semiconductors and computers. Higher precision of the instruments leads to greater impact on measurement results when affected by abnormal factors. Therefore, real-time anomaly detection for precision instrument measurements is crucial. This paper uses the robust random cut forest (RRCF) algorithm as a basis to explore its application in the anomaly detection of measurement results of precision instruments. The model is constructed using the RRCF algorithm, and the anomaly score is returned by the degree of change in model complexity after inserting sample points. The algorithm has better detection performance and operational efficiency for real-time data flow, this paper applies it to real-time anomaly detection of precision instrument measurement results, and compares its indexes such as accuracy, precision and recall with other algorithms, and the results show that all the indexes have been improved.