Medium-term electric energy demand forecasting using Nadaraya-Watson estimator

Abstract

This work is focused on monthly electricity demand prediction, which is necessary for the maintenance planning in power systems as well as for negotiation forward contracts. In the proposed approach patterns of the load time series are defined, which unify input and output data and filter out the trend. Relationships between inputs and outputs simplified due to patterns are modeled using nonparametric regression: Nadaraya-Watson estimator. In the experimental part of the work the model is examined using real-world data. The results are encouraging and confirm the high accuracy of the model and its competitiveness compared to other forecasting models.