Electricity Load Forecasting Using an Ensemble of Optimally-Pruned and Basic Extreme Learning Machines
Lufuno Marwala, Bhekisipho Twala · 2017 9th IEEE-GCC Conference and Exhibition (GCCCE) · 2017
In this paper the application of artificial intelligence on one-step ahead forecasting of electricity consumption is investigated. The total electricity consumption data sampled on a monthly basis (monthly consumption) from January 1985 to December 2011 in South Africa is used. Neural networks, Neuro-fuzzy systems, support vector machines and optimally-pruned and basic extreme learning machines (ELM) were used to develop nonlinear ensemble models for forecasting and their performance is compared. It was found that extreme learning machines significantly outperform traditional techniques except for support vector machines.