| Title |
A Study on the GAN Algorithm Performance Improvement Method in Motor Failure Diagnosis Using Deep Learning Algorithm |
| Authors |
한지훈(Ji-Hoon Han) ; 최동진(Dong-Jin Choi) ; 박상욱(Sang-Uk Park) ; 홍선기(Sun-Ki Hong) |
| DOI |
https://doi.org/10.5370/KIEE.2020.69.11.1732 |
| Keywords |
Motor fault diagnosis; Deep learning; GAN; Lack of the data; DT-CNN |
| Abstract |
When deep learning is applied to motor fault diagnosis, the Generative Adversarial Network (GAN) algorithm is used to compensate for the insufficient number of data. However, the model’s performance includes not only classification performance, but also overfitting degree and outlier detection performance. Unlike classification performance, outlier detection performance is affected not only by the amount of data but also by the quality. However, increasing similar data is not the best way to increase the model’s overall performance. Therefore, an RMSE-based data evaluation technique is proposed to find virtual data that maximizes model performance. It was confirmed that the proposed method helps to improve the outlier data detection performance of the model. |