Transactions of FAMENA, Vol. 47 No. 1, 2023.
Original scientific paper
https://doi.org/10.21278/TOF.471040722
The Matthew Effect of a Fault Classification Mechanism and Its Application
Shihua Wang
; School of Science, Guangdong University of Petrochemical TechnologyMaoming, China
Shaolin Hu
; School of Automation, Guangdong University of Petrochemical Technology, Maoming, China
Qinghua Zhang
; School of Automation, Guangdong University of Petrochemical Technology, Maoming, China
Abstract
When using the classification algorithm to classify a single sample, the classification accuracy often cannot achieve an ideal effect. To solve this problem, the following two aspects of research work are carried out and presented in this paper. On the one hand, according to the memory characteristics of mechanical faults, a voting classification mechanism for the sample sequence to be classified is proposed. It is found that the classification mechanism of the sample sequence to be classified with memory has the Matthew effect of accumulated advantage. Using this effect, one can improve the accuracy of fault classification. On the other hand, because the length of the sample sequence to be classified increases, the delay of the classification results increases. To solve this problem, the classification algorithm is optimized to minimize the delay on the assumption that the classification accuracy meets the expected requirements.
Keywords
fault diagnosis; machine learning; the Matthew effect
Hrčak ID:
299475
URI
Publication date:
19.4.2023.
Visits: 895 *