Predicting Building Energy Consumption from Electricity Meters Using Prophet: A Computational Study with Nested Cross-Validation on Benchmark Time Series Data

International Symposium on Innovative Approaches in Smart Technologies (ISAS), pp. 1–6

Abstract

In this study, we investigate the application of Prophet, an open-source forecasting tool, to predict electricity energy consumption using a substantial benchmark dataset. The considered benchmark dataset includes time series data from electricity meters in hourly resolution from 200 buildings. Accurate energy usage prediction is essential for efficient energy management and sustainability. Our computational analysis showcases the capabilities and limitations of Prophet in capturing the intricate patterns in energy consumption data. Through a comprehensive evaluation employing nested cross-validation for time series, we highlight the model’s performance.

Authors 2

1. Alina Shcherbakova Aachen Faculty of Electrical Engineering and Information Theory

   FH Aachen

   Affiliation as printed

   Aachen University of Applied Sciences,Faculty of Electrical Engineering and Information Theory,Aachen,Germany

2. Lena Charlotte Altherr Aachen Institute of Smart Building Engineering Aachen University of Applied Sciences Faculty of Electrical Engineering and Information Theory

   FH Aachen

   Affiliation as printed

   Institute of Smart Building Engineering Aachen University of Applied Sciences,Faculty of Electrical Engineering and Information Theory,Aachen,Germany