Analysis of Unsupervised Consumption Anomaly Detection in Sports Facilities using Artificial Intelligence-Based Data Analytics: A Case Study

Abstract

Sports facilities have exceptionally high energy demand due to the extensive operational requirements and high-occupancy seasonal rates. Towards promoting efficient energy usage and minimal losses, consumption anomaly detection in sports facilities is addressed in this work using Artificial intelligence (AI)-based analytics approaches. Traditional AI-based data analytics approaches are applied in a practical context for a local sports complex. The actual unlabeled operation data of the facility are used and a case-specific comparative analysis of the various approaches is presented where AI-based data labeling is used. The characteristics of the different algorithms are contextually discussed. It was found that the size and distribution of the training datasets influence the performance of the different algorithms. This study represents preliminary findings on the topic with a promising potential for further research.