An empirical comparison of cross-impact models for forecasting sales |
| |
Affiliation: | 1. Department of Operations Management, Copenhagen Business School, Copenhagen, Denmark;2. School of Social Sciences, Sodertorn University, SE-141 89 Stockholm, Sweden;3. School of Innovation, Design and Engineering, Mälardalen University, Box 325, 631 05 Eskilstuna, Sweden;4. School of Business, Maynooth University, Maynooth, Co. Kildare, Ireland;1. Room 3042, No. 9 Student Building, Yuquan Campus, Zhejiang University, Hangzhou 310027, China;2. Carnegie Mellon University, 444 Washington Blvd., Apt. 1517, Jersey City, NJ 07310, United States;3. No. 11 Classroom Building, Yuquan Campus, Zhejiang University, Hangzhou 310027, China. |
| |
Abstract: | This paper compares a set of four cross-impact models: (1) additive, (2) likelihood multiplier, (3) R-space, and (4) a model constructed by the author. This is done by examining a forecasting problem encountered by an industrial firm. The forecasting problem was to study the market trend in order to decide whether to expand the production capacity of a ceramics plant. In spite of their different theoretical premises, the models yielded similar results. However, only the R-space model produced results that differed from the others. The paper also suggests a method that should avoid some internal contradictions of the cross-impact models. |
| |
Keywords: | |
本文献已被 ScienceDirect 等数据库收录! |
|