Current Result Document : 62 / 62
| 한글제목(Korean Title) |
온라인 소매업체 데이터를 이용한 특정 패션 스토어의 시계열 판매량 예측을 위한 딥러닝 애플리케이션 |
| 영문제목(English Title) |
A Deep Learning Application for Forecasting the Time-Series Sales Volume of a Specific Fashion Store Using Online Retailer Data |
| 저자(Author) |
라젠드란 무랄이다란
홍봉희
Rajendran Muralidharan
Hong Bonghee
|
| 원문수록처(Citation) |
VOL 49 NO. 02 PP. 0292 ~ 0294 (2022. 12) |
한글내용
(Korean Abstract) |
|
영문내용
(English Abstract) |
Predicting customer preferences is a challenging task in the fashion industry due to the rapid changes in the market and the new launches of every season. The classical forecasting methods and traditional machine learning models cannot resolve this problem as the new apparel products lack historical sales data. In this paper, we propose a novel image-based multi-model transformer consisting of encoding-decoding modules to capture the trends among interrelated attributes with specified embedding layers for heterogeneous data. We evaluate our approach on real-world data collected from NineOunce, a Korean fashion company. Our dataset is fabricated with more than 15k products sold from 2017 which consists of product images, historical sales data, and some valuable external factors like season, sales price, and discount. Additionally, we also investigate the importance of including the knowledge of a familiar Korean e-tailer NAVER Shopping, which is filtered with the fashion apparel products category. Finally, our approach highly elevates the accuracy of the forecast by using the knowledge of exogenous information, compared to the other existing benchmark models. |
| 키워드(Keyword) |
|
| 파일첨부 |
PDF 다운로드
|
