Forecasting innovation diffusion of products using trend-weighted fuzzy time-series model

The time-series models have been used to make reasonably accurate predictions in weather forecasting, academic enrolment, stock price, etc. This study proposes a novel method that incorporates trend-weighting into the fuzzy time-series models advanced by Chen’s and Yu’s method to explore the extent to which the innovation diffusion of ICT products could be adequately described by the proposed procedure. To verify the proposed procedure, the actual DSL (digital subscriber line) data in Taiwan is illustrated, and this study evaluates the accuracy of the proposed procedure by comparing with different innovation diffusion models: Bass model, Logistic model and Dynamic model. The results show that the proposed procedure surpasses the methods listed in terms of accuracy and SSE (Sum of Squares Error).