Skip to Main Content
This paper introduces several soft models such as the genetic algorithm, neural network and fuzzy regression to study and predict the Internet growth in several OECD nations. First a linear version of an augmented diffusion model is designed. The augmented diffusion model is designed by including the impact of an economic indicator, GDP per capita in to the model. In the next stage the soft models are built, using the augmented diffusion model as the base model. Performance and forecasting measures from these soft models show that these soft models provide improvements over the augmented diffusion model. We also discuss how the information from the models can be reused and integrated.
Date of Conference: 8-10 Nov. 2004