Self-Organizing Methods in Modeling: Gmdh Type Algorithms
Self-Organizing Methods in Modeling: Gmdh Type Algorithms
International Journal of Intelligent Systems
Data mining via rules extracted from GMDH: an application to predict churn in bank credit cards
KES'10 Proceedings of the 14th international conference on Knowledge-based and intelligent information and engineering systems: Part I
Kernel group method of data handling: application to regression problems
SEMCCO'12 Proceedings of the Third international conference on Swarm, Evolutionary, and Memetic Computing
Information Sciences: an International Journal
A methodological approach to mining and simulating data in complex information systems
Intelligent Data Analysis
Hi-index | 0.01 |
The electric power industry is in transition as it moves towards a competitive and deregulated environment. In this emerging market, traditional electric utilities as well as energy traders, power pools and independent system operators (ISOs) need the capability to predict as precisely as possible how much energy their customers will use in the near future. This paper presents a medium-term energy demand forecasting system that helps utilities identify and forecast energy demand for each of the end-use consumption sector of the energy system, representing residential, industrial, commercial, non-industrial, entertainment and public lighting load. The demand forecasting system is organized and implemented in a modular fashion using high accuracy forecast models. These models are developed for each sector to account for the growth trends and seasonal effects. A comparative evaluation of various traditional and neural network-based methods for obtaining the forecast of monthly energy demand was carried out. Among the models tested, artificial neural network (ANN)-based models were determined to produce better results. In particular, group method of data handling (GMDH) neural network, composed of self-organizing active neurons, was proven very effective in producing forecasts that were significantly more accurate and less labor-intensive than traditional time-series and regression-based models.