Combining various dissimilarity measures for clustering electricity market prices

Abstract

The analysis of electricity markets of the European countries aims to better understand their degree of integration, which is relevant for the development of an internal market of electricity in the European Union. This study resorts to clustering of time series of hourly prices of electricity (in e/MWh) observed in the day-ahead market in 2018. The proposed approach relies on the combination of different dissimilarity measures which can capture differences in time series trends, (prices) values, cyclical behaviors and autocorrelation patterns. The results obtained, enable to provide some insights on the role of the different dissimilarity measures in the clustering process. Furthermore, they provide a clustering solution with coherent substantive interpretation and, interestingly, one that reveals the natural patterns of geographic proximity.