[국제학술대회] Abnormal Data Classification Based on SSA-AELSTM
- 김동현
- 조회수1103
- 2021-10-14
-Title: Abnormal Data Classification Based on SSA-AELSTM
-Journal/Conference: International Conference on Computer Science and Computational Intelligence 2021 (ICCSCI 2021), pp. xx-yy, November 2021
-Authors: Donghyun Kim, Seokju Oh and Jongpil Jeong
-DOI :
-Journal/Conference Link: https://socs.binus.ac.id/iccsci
Abstract:
Smart factories and big data are important factors in the 4th industrial revolution. Smart factories aim at automation and integration, but the most important part is the use of data. Although many studies are underway on the maintenance and quality management of big data-based production equipment, facility data collected for industrial data analysis has more normal data than abnormal data. In addition, a lot of energy is consumed in the data preprocessing process to analyze the collected data. Therefore, to maintain production equipment and manage quality, data classification technology that is easy for data analysis by classifying abnormal data only with normal data is needed. In this paper, we propose the classification of outliers in target data of univariate cycle data collected from production facilities through SSA-AELSTM. SSA-AELSTM is a hybrid technique that combines SSA techniques that are effective in reducing noise in time series data and LSTM automatic encoder that showed excellent performance in time series outlier detection.
-Status: Accept (2021/10/14)