Electric load forecasting with recency effect: A big data approach

TLDR: A comprehensive study to model the recency effect using a big data approach and two interesting findings are presented: 1) the naive models are not useful for benchmark purposes in load forecasting at aggregated level due to their lack of accuracy; and 2) slicing the data into 24 pieces to develop one model for each hour is not necessarily better than building one interaction regression model using all 24 hours together.

About: This article is published in International Journal of Forecasting.The article was published on 2016-07-01 and is currently open access. It has received 167 citations till now. The article focuses on the topics: Linear model & Moving average.

Figures

Figure 7. Heat map of MAPE values (in %) for recency effect modeling based on test data (2007) of Z21.

Table 2 Different implementations of recency effect modeling

Table 4 MAPE values (in %) of the Recency effect vs. benchmark models on 21 zones of the GEFCom2012 load forecasting data (24-hour ahead ex post forecasting for 2007). Z4 experienced a major outage; Z9 is an industrial customer.

Table 5 lists the MAPE values of the recency effect model (option A), two advanced models mentioned above, and the benchmark model. This time we zoom into the error statistics by hour. We only show the results from Z21 to avoid verbose presentation, because results from the other zones do not add additional

Table 5 lists the MAPE values of the recency effect model (option A), two advanced models mentioned above, and the benchmark model. This time we zoom into the error statistics by hour. We only show the results from Z21 to avoid verbose presentation, because results from the other zones do not add additional

Table 4 MAPE values (in %) of the Recency effect vs. benchmark models on 21 zones of the GEFCom2012 load forecasting data (24-hour ahead ex post forecasting for 2007). Z4 experienced a major outage; Z9 is an industrial customer.

Frequently asked questions

Q1. What have the authors contributed in "Electric load forecasting with recency effect: a big data approach" ?

In this paper, the authors present a comprehensive study on modeling recency effect through a big data approach. Using the case study based on data from the load forecasting track of the Global Energy Forecasting Competition 2012, the authors first demonstrate that a model with recency effect outperforms its counterpart ( a. k. a., Tao ’ s Vanilla Benchmark Model ) in forecasting the load series at the top ( aggregated ) level by 18 % to 21 %. The authors then apply recency effect modeling to customize load forecasting models at low level of a geographic hierarchy, again showing the superiority over the benchmark model by 12 % to 15 % on average. Finally, the authors discuss four different implementations of the recency effect modeling by hour of a day. The authors take advantage of the modern computing power to answer a fundamental question: how many lagged hourly temperatures and/or moving average temperatures are needed in a regression model to fully capture recency effect without compromising the forecasting accuracy ?