| 학회 |
한국화학공학회 |
| 학술대회 |
2017년 봄 (04/26 ~ 04/28, ICC 제주) |
| 권호 |
23권 1호, p.190 |
| 발표분야 |
공정시스템 |
| 제목 |
Driving mode estimation under concept drift by ensemble learning |
| 초록 |
Estimating the driving mode for individual drivers through real-time operational data is crucial to post-treatment of the exhaust gases. However, current methods do not estimate the driving mode for each driver, and ignore the influence of periodic or occasional disturbance. It is referred to as concept drift problem. In this study, we explain how to connect real-time operation data and driving mode using ensemble learning to cope with concept drift. Through simulation based on real data of Hyundai motors, we compare the performance of the fixed model in various concept drift situations and the driving mode classification performance of the proposed method. Compared with existing fixed models, estimates from ensemble learning show better driving mode classification and these results enable better post-treatment of the exhaust gases. |
| 저자 |
정동휘1, 이종민1, 임산하1, 정창호2, 김창환2, 김용화2
|
| 소속 |
1서울대, 2현대자동차 |
| 키워드 |
공정모델링 |
| E-Mail |
|
| 원문파일 |
초록 보기 |
