A novel assessable data augmentation method for mechanical fault diagnosis under noisy labels. (July 2022)

• Record Type: Journal Article
• Authors: Zhang, Xin; Wu, Bo; Zhang, Xi; Zhou, Quan; Hu, Youmin; Liu, Jie
• Is Part Of: Measurement. Volume 198(2022)
• Publication Date: 2022-07
• Subjects: Data augmentation -- Generative adversarial network -- Fault diagnosis -- Noisy label; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2022.111114
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: We proposed a novel assessable data augmentation named ADA, which can improve the quality of generated samples under noisy labels. The ADA method can be mainly divided into two parts: sample quality assessment (SQA) and sample generation based on WGAN-gp. The proposed SQA can give quantitative descriptions of data quality for all training samples. A statistic tool named influence function (IF) was introduced into SQA, which can accelerate the process of sample assessment. The experimental results demonstrate that ADA can effectively improve the fault diagnosis accuracy for various classifiers. Abstract: Data augmentation technology has achieved great success to expand the training set for several years. As a representative technology, generative adversarial network and its variants are widely applied in many data augmentation tasks. But the quality of training samples is rarely considered. In this paper, a novel assessable data augmentation named ADA is proposed for mechanical fault diagnosis under noisy labels. First, a sample quality assessment procedure including assessment model construction, approximate calculation based on influence function and screening decision is presented. Thereby, the optimized training set can be obtained. Then, the WGAN-gp model can be established based on the optimized training set and the data augmentation can be accomplished. Finally, a classifier can be trained with the expanded training set and achieve the task of fault diagnosis. The results of two experiments show that the proposed ADA method can effectively improve the fault diagnosis accuracy for various classifiers.