2020/12/14 by Yamila Barrera, Leonardo Boechi, Barrera, Yamila +13
Computer Science · Economics, Econometrics and Finance · #Applications (stat.AP) #Complex Systems and Time Series Analysis #Data Management and Algorithms #FOS: Computer and information sciences #Machine Learning (cs.LG) #Time Series Analysis and Forecasting
paper · pdf · doi:10.48550/arxiv.2012.07487
openalex publication_date 2020/12/14 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
The Reseau de Transport d'Electricit 'e (RTE) is the French main electricity\nnetwork operational manager and dedicates large number of resources and efforts\ntowards understanding climate time series data. We discuss here the problem and\nthe methodology of grouping and selecting representatives of possible climate\nscenarios among a large number of climate simulations provided by RTE. The data\nused is composed of temperature times series for 200 different possible\nscenarios on a grid of geographical locations in France. These should be\nclustered in order to detect common patterns regarding temperatures curves and\nhelp to choose representative scenarios for network simulations, which in turn\ncan be used for energy optimisation. We first show that the choice of the\ndistance used for the clustering has a strong impact on the meaning of the\nresults: depending on the type of distance used, either spatial or temporal\npatterns prevail. Then we discuss the difficulty of fine-tuning the distance\nchoice (combined with a dimension reduction procedure) and we propose a\nmethodology based on a carefully designed index.\n