Probabilistic Hourly Load Forecasting Using Additive Quantile Regression Models

Abstract

Short-term hourly load forecasting in South Africa using additive quantile regression (AQR) models is discussed in this study. The modelling approach allows for easy interpretability and accounting for residual autocorrelation in the joint modelling of hourly electricity data. A comparative analysis is done using generalised additive models (GAMs). In both modelling frameworks, variable selection is done using least absolute shrinkage and selection operator (Lasso) via hierarchical interactions. Four models considered are GAMs and AQR models with and without interactions, respectively. The AQR model with pairwise interactions was found to be the best fitting model. The forecasts from the four models were then combined using an algorithm based on the pinball loss (convex combination model) and also using quantile regression averaging (QRA). The AQR model with interactions was then compared with the convex combination and QRA models and the QRA model gave the most accurate forecasts. Except for the AQR model with interactions, the other two models (convex combination model and QRA model) gave prediction interval coverage probabilities that were valid for the 90%, 95% and the 99% prediction intervals. The QRA model had the smallest prediction interval normalised average width and prediction interval normalised average deviation. The modelling framework discussed in this paper has established that going beyond summary performance statistics in forecasting has merit as it gives more insight into the developed forecasting models.

Averaging Predictive Distributions Across Calibration Windows for Day-Ahead Electricity Price Forecasting

“…In this paper, we introduce two extensions of the averaging across calibration windows concept to probabilistic forecasting. The first one is based on Quantile Regression Averaging (QRA) of Nowotarski and Weron [7], which has been found to perform very well in several test cases, including electricity price [3,11,23,24], load [25][26][27] and wind power forecasting [28]. Recall, that QRA involves applying quantile regression [9] to a pool of point forecasts.…”

Short-term Forecasting of Electricity Consumption in Buildings for Efficient and Optimal Distributed Energy Management

“…The RegPSO algorithm considers the particles in the swarm are very close to each other (premature convergence or stagnation occurs) and initiates the regrouping of the particles when (16) is satisfied. When the stagnation is detected by the condition specified in (16), regrouping of the particles in the swarm is initialized. The regrouping performs in the search space about the center of the global best position, Gbest.…”

Point and Interval Forecasting of Zonal Electricity Prices and Demand Using Heteroscedastic Models: The IPEX Case

“…To obtain them, he proposed an additive non-parametric model whose location, scale and shape parameters were non-linear additive functions of the covariates. Additive models for conditional expectation were proposed by [42,43] and extended to predict individual quantiles by [44,45]. Linear quantile regression models have also been used by [46], for day-ahead and intra-daily markets, to produce probabilistic forecasts of hourly spot prices with the aim of reducing forecast bias and the width of forecast intervals.…”