A Synthesis of Monte Carlo and Noise-in-Variable Model Techniques for Energy Demand Modelling

Abstract

This paper considers the issue of uncertainty analysis in energy demand modelling for cases where the data is noisy. This situation is particularly relevant to the developing countries energy system where there is significant measurement error in sodo-economic variables. We present a probabilistic approach which is a combination of noise-in-variable and Monte Carlo simulation techniques as a forecasting procedure for such cases. A simple residential electricity demand model for Nigeria is analysed to illustrate the proposed approach.