A fine-grained multi-category imbalanced fault classification method based on knowledge distillation

Abstract

The invention discloses a fine-grained multi-category unbalanced fault classification method based on knowledge distillation. The method combines algorithms such as knowledge distillation and hierarchical clustering to solve the multi-category unbalanced fault classification problem. For the multi-class fault classification problem, the knowledge distillation method is first used for fault classification. Aiming at the characteristics of similarity between samples of homogeneous categories and large differences between samples of heterogeneous categories in the imbalance problem, a hierarchical clustering method is used to cluster all category samples according to the clustering results of the category center points. class, so as to obtain fine-grained cluster classes. Finally, fine-grained fault classification is performed for each cluster class. For each cluster class, a student network will be established, and finally spliced ​​to optimize the multi-student network together. Under the guidance of the global information of the teacher's network, combined with the fine-grained fault classification of the multi-student network. Compared with other existing methods, the method of the invention has good effect and applicability.

Description

technical field [0001] The invention belongs to the field of industrial process monitoring, and in particular relates to a fine-grained multi-category unbalanced fault classification method based on knowledge distillation. Background technique [0002] In machine learning or deep learning classification, the imbalance of the number of class samples is a very common problem, which exists in various fields, such as bioinformatics, smart grid, medical imaging, fault diagnosis. Most existing classification methods are based on the assumption that the underlying distribution of observed data is relatively balanced. However, real industrial datasets usually violate this assumption and present a skewed distribution or even an extremely unbalanced distribution of class sample sizes. For example, data-driven fault classification is an important part of industrial process monitoring, which exhibit unbalanced skewed distributions due to different frequencies of fault occurrence. In t...

AI Technical Summary

Problems solved by technology

However, real industrial datasets often violate this assumption and exhibit skewed distributions or even extremely unbalanced distributions of class sample sizes

Images