| 제목 | 온라인 리테일러에서 수요 패턴 예측을 위한 컨볼류션 LSTM 적용 방안 |
|---|---|
| Title | Application of Convolutional LSTM for Predicting Demand Patterns in Online Retailer |
| 저자 |
정재원* (한양대학교 산업공학과) 임성태 (한양대학교 산업공학과) 이승민 (한양대학교 산업공학과) 공형준 (한양대학교 산업공학과) |
| Author |
J. Jeong†(Department of Industrial Engineering, Hanyang Universi) S. Lim(Department of Industrial Engineering, Hanyang Universi) S. Lee(Department of Industrial Engineering, Hanyang Universi) H. Kong(Department of Industrial Engineering, Hanyang Universi) |
| Bibliography | Journal of Logistics Science & Technology, 1(1),13~28, 2020, |
| DOI | |
| Key Words | Demand pattern forecasting, On-line retailing, Convolutional LSTM, Customer orders, Order management |
| Abstract | Demand pattern forecasting would be surely one of significant modules for successful and profitable on-line retail business. Our work studies the mechanism for demand pattern forecasting at the on-line retailers in a context of e-commerce business. This paper deals with demand pattern forecasting on a basis of time series prediction. We review papers about demand forecasting in a lot of application areas and suggest a new approach to time series prediction using Convolutional LSTM by generating the images for corresponding demand patterns. And it is observed that our proposed approach could elicit more accurate results than the current existing techniques such as seasonal ARIMA and Holt-Winters Technique. Numerical studies using the data for an e-commerce company have been done in order to validate the performance of the suggested demand pattern forecasting mechanism using the Convolutional LSTM. Finally, we provide the meaningful measures ensuring the usefulness and applicability of the proposed demand pattern forecasting mechanism. |
| PDF download | JM_1[1]-P.pdf |