Calibration window selection based on change-point detection for forecasting electricity prices

TL;DR

This paper introduces a novel calibration window selection method for electricity price forecasting using change-point detection, leading to improved accuracy over traditional approaches.

Contribution

It applies the Narrowest-Over-Threshold change-point detection to select relevant subperiods for autoregressive modeling, enhancing forecast precision.

Findings

Significant improvement in forecasting accuracy

Outperforms traditional methods like ARHNN

Effective on German EPEX SPOT electricity prices

Abstract

We employ a recently proposed change-point detection algorithm, the Narrowest-Over-Threshold (NOT) method, to select subperiods of past observations that are similar to the currently recorded values. Then, contrarily to the traditional time series approach in which the most recent $\tau$ observations are taken as the calibration sample, we estimate autoregressive models only for data in these subperiods. We illustrate our approach using a challenging dataset - day-ahead electricity prices in the German EPEX SPOT market - and observe a significant improvement in forecasting accuracy compared to commonly used approaches, including the Autoregressive Hybrid Nearest Neighbors (ARHNN) method.