Two simple models for the early and middle stage prediction of innovation diffusion

When a technological innovation, a new technology or a new technological product is introduced in an economic sector, a market, a country, or in the international environment, the technology manager, based on the already observed data, needs, from the early stages of the diffusion process, to make reliable predictions for the future trends. In practice, forecasts are frequently based on the logistic or on the various sigmoid-form models after an estimation of the upper limit, a case which cannot always be met in the early stages of the diffusion. This work tries to improve the predictions based on early and middle stage data by introducing two new innovation diffusion models. Both models arise from the generalized rational models (GRM) and may be useful for early and middle stage predictions because they do not include a parameter expressing the upper limit. The properties of these models are studied and the methodology of the used nonlinear regression analysis techniques is discussed. Two applications of the proposed models on time-series data selected from the technological innovation diffusion area indicate the usefulness of these models. Finally an extension of one of the models and a comparative application follows.