Created by W.Langdon from gp-bibliography.bib Revision:1.7428
The objective of the dissertation is to propose, using quantitative forecast methods, a new Empirical deterministic Model (EM) or Forecasting Model (FM) for short-term forecast of the cumulative broadband adoption of a new technology and a hybrid Genetic Programming method (hGP) in fitting and forecasting of the broadband penetration data.
Initially, the fitting and forecasting performance of the EM (FM) is compared with some widely used diffusion models for the cumulative adoption of new telecommunication products, namely, Logistic, Gompertz, Flexible Logistic (FLOG), Box-Cox, Richards and Bass model. The fitting and forecasting processes are implemented in broadband penetration official data for Greece and the EM yields well enough statistical indices.
The introduction of the hybrid Genetic Programming method (hGP) in fitting and forecasting of the broadband penetration data follows. Some well-known diffusion models, such as those of Gompertz, Logistic and Bass, are embedded in the hGP initial population of the solutions to accelerate the algorithm. The produced solutions-models of the hGP are used in fitting and forecasting the adoption of broadband penetration in Organisation for Economic Co-operation and Development (OECD) countries. The results of the optimised diffusion models are compared to those of the hGP. The hGP generates solutions with high performance statistical indices and the solutions cooperate with the existing diffusion models, thus allowing multiple approaches to forecasting.
Finally, a modified hGP method in forecasting constitutes an expansion of the hGP. It has been improved by the introduction in the initial population of some diffusion models variation for long-term technological forecasting purposes, such as Bi-Logistic and LogInLog. New rules for initial populations production and models in the functions set are inserted. The co-operation of the method with macroeconomic indicators such as Gross Domestic Product per Capita (GDPpC) and Consumer Prices Index (CPI), leads to the creation of forecasting models and scenarios for medium and long-term level of predictability. The modified hGP has been implemented in the datasets of the mobile subscribers and fixed broadband penetration in OECD countries and it achieves well enough forecasting performance.
In conclusion it should be emphasised that the development of the new modified method of Evolutionary Algorithms can be applied to other scientific fields as well.",
Genetic Programming entries for Konstantinos Salpasaranis