Forecasting of Mobile Subscriptions in Asia Pacific Using Bass Diffusion Model

In today's dynamic world, mobile communication has changed the lifestyle of many people. Forecasting of mobile subscriptions in a country is one of the research areas. In this paper, Bass diffusion model is used to forecast the number of mobile service subscribers in major countries in Asia Pacific. Technically, we compare two forecasting methods: Bass diffusion model and diffusion by analogy. One is popular in research and the other is commonly applied in practice. Two estimation methods for Bass diffusion model are also compared: adaptive nonlinear least square (adaptive NLS) and genetic algorithm (GA). The results show that Bass diffusion model in general performs better than diffusion by analogy based on the in-sample sum of squared errors (SSE) and out-of-sample SSE. On the other hand, adaptive NLS and genetic algorithms are comparable in generating reasonably sound forecasting results