Forecasting the OECD Fixed Broadband Penetration with Genetic Programming Method, Diffusion Models and Macro-Economic Indicators

Abstract This paper presents the implementation of a modified Genetic Programming (GP) method in forecasting fixed broadband telecommunications penetration percentage in Organisation for Economic Co-operation and Development (OECD) countries. The specific GP method combines the use of known diffusion models for technological forecasting purposes, such as Logistic, Gompertz and Bass and the GP. The combination method produces both time dependant and causal models with high performance statistical indicators. Also, multiple approaches to forecasting can be implemented, mainly with no big datasets.