An empirical study of intelligent expert systems on forecasting of fashion color trend

Yong Yu, Chi Leung Hui, Tsan Ming Choi

Research output: Journal article publicationJournal articleAcademic researchpeer-review

31 Citations (Scopus)

Abstract

Forecasting future color trend is a crucially important and challenging task in the fashion industry including design, production and sales. In particular, the trend of fashion color is highly volatile. Without advanced methods, it is very hard to make fashion color trend forecasting with reasonably high accuracy, and it is a handicap for development of the intelligent expert systems in fashion industry. As a result, many prior works have employed traditional regression models like ARIMA or intelligent models such as artificial neural network (ANN) and grey model (GM) for conducting color trend forecasting. However, the reported accuracies of these forecasting methods vary a lot, and there are controversies in the literature on these models' performances. As a result, in this paper, we systematically compare the performances of ARIMA, ANN and GM models and their extended family methods. With real data analysis, our results show that the ANN family models, especially for Extreme Learning Machine (ELM) with Grey Relational Analysis (GRA), outperform the other models for forecasting fashion color trend.
Original languageEnglish
Pages (from-to)4383-4389
Number of pages7
JournalExpert Systems with Applications
Volume39
Issue number4
DOIs
Publication statusPublished - 1 Mar 2012

Keywords

  • ARIMA
  • Artificial neural network
  • Color trend
  • Fashion design
  • Forecasting
  • Grey model

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications
  • Engineering(all)

Cite this