The ITISE (International conference on Time Series and Forecasting) is an annual three-day conference, which was held for the ninth time in 2023 and took place over the period 12.07.-14.07.2023 in Meloneras/Gran Canaria. The organiser of the conference is the Universidad de Granada. The aim of the conference is to provide a discussion forum for the latest ideas and findings in the areas of foundations, theory, models and applications for interdisciplinary research in the field of time series analysis and forecasting. ITISE thus addresses a wide range of disciplines, from computer science, mathematics and statistics to applications in engineering, medicine and econometrics. In this way, ITISE creates a plurality of ideas from different subject areas and thus strengthens the exchange across the disciplines, which show great differences in their applications but only minor differences in the methods.
At ITISE 2023, the paper "Increasing the Performance and Plausibility of Machine Learning via Data Analysis Techniques" was presented by the authors of the Department of Energy Efficiency Optimisation Silas Aaron Selzer, Fabian Bauer and Prof. Peter Bretschneider and was awarded the Best Paper Award at the conference dinner. The paper compares different filtering methods as a subset of data analysis techniques for feature selection. Using examples from power engineering, the strengths of data analysis as a pre-processing step and its effectiveness are demonstrated. The use of data analysis techniques not only increases the performance of machine learning methods, but also their plausibility. This is achieved by identifying the significant input variables and reducing model complexity, which in turn leads to trustworthy and reliable models. The paper is also published as a chapter in the book "Contributions to Statistics" by Springer Verlag.
The work was made possible by the cooperation with 50Hertz Transmission GmbH, which provided the data for the examples (calculation of the conductor rope temperature, the grid losses and the vertical grid load). The authors would like to take this opportunity to thank them again for providing the data and for the constructive discussions.
Source: Silas Aaron Selzer, Department of Energy Use Optimisation