Collaborative Forecasting Models for the Machine Tools Industry via the Internet

The purpose of this research is to investigate appropriate collaborative forecasting models, both for acceptable accuracy and effectiveness for information sharing via the Internet to partners. Several forecasting models were investigated in this research. For comparing the models, five years historical data are collected from a lathe machine manufacturer in Taiwan, and one of its partners – a ball screw vendors in. After data verification, we conclude that the MLR model is the best forecasting model among those compared. Moreover, models with multiple variables, contributed from vendors and agencies, have better performance than models with a single variable. The conclusion in this research could be a reference for forecasting model creation in Collaborative Planning, Forecasting and Replenishment (CPFR) via the Internet.