A0388
Title: Robust functional principal component analysis for detecting anomalous behaviors in electricity markets
Authors: Luigi Grossi - University of Parma (Italy) [presenting]
Andrea Cerasa - Joint Research Centre (Italy)
Mara Sabina Bernardi - European Commission - Joint Research Center (Italy)
Fany Nan - Joint Research Center of the European Commission Ispra (Italy)
Abstract: The main objective is to propose a procedure for detecting outliers that may be linked to market manipulation, utilizing microdata from daily auctions that establish the equilibrium price and quantity in deregulated electricity markets. In the context of electricity markets, outlier detection methods can be applied to identify abnormal price movements deviating from typical market behavior. However, establishing a direct link is not always easy, as observed outlying prices may be influenced by other factors whose combined effects are difficult to disentangle. In contrast to previous approaches, observing and approximating the shape of the supply curves is advocated instead of focusing solely on observed prices. Through techniques like functional principal component analysis (FPCA), functional data analysis allows for the identification of key patterns and anomalies in the shape of the supply curve that may indicate illegal actions. The detection of potential outliers is based on the residuals obtained from a robust version of the functional principal component analysis (FPCA) model. This shift is motivated by the fact that market manipulation events involve the strategic decisions of market operators, directly influencing the shape of the supply curve. Therefore, the fundamental hypothesis is that the strategic behavior of operators is established assuming a non-flexible demand curve.
