Short-term Load Forecasting for Special Days Using Rule-based Models: A Case Study for France

Arora, Siddharth and Taylor, James (2018) Short-term Load Forecasting for Special Days Using Rule-based Models: A Case Study for France. European Journal of Operational Research. (Accepted)

Full text not available from this repository.


This paper presents a case study on short-term load forecasting for France, with emphasis on special days, such as public and national holidays. Existing methods for load forecasting focus mainly on normal working days, while special days are often ignored during the modeling process. We generate short-term load forecasts for normal and special days in a coherent and unified framework, using a rule-based methodology. The proposed methodology encapsulates prior domain knowledge of load profiles into the statistical model. In addition to special day effects, we accommodate the intraday, intraweek, and intrayear seasonality in load, whereby we consider the original version and corresponding rule-based triple seasonal adaptation of HoltWinters-Taylor (HWT) exponential smoothing, seasonal autoregressive moving average (SARMA), artificial neural networks (ANNs), along with intraday and intraweek singular value decomposition (SVD) based exponential smoothing methods. Using nine years of half-hourly load for France, we evaluate point and density forecasts, for lead times ranging from one halfhour up to a day ahead. Overall, the rule-based SARMA method generated the most accurate point and density forecasts.

Item Type: Article
Keywords: management science, forecasting
Subject(s): Management science
Date Deposited: 27 Nov 2017 10:02
Last Modified: 16 Jan 2018 16:04
Funders: N/A

Actions (login required)

Edit View Edit View