In smart buildings, efficient energy consumption is one of the biggest challenges to solve, which can contribute to reduce the global warming of our planet, due to its relevance. In this paper, a time series to image (TSI) based model is introduced to identify anomalous energy consumption in residential buildings. It has a novel encoding scheme to transform univariate time series data into images for extracting the useful information and one-class support vector machine (OCSVM) for the classification task. The TSI extracts descriptive and representative feature spaces from the data and encode them into images using Markov Transition Function (MTF). We empirically evaluate the proposed model over publicly available real world dataset and compared the results with the state-of-the-art method. The obtained results are competitive and confirm the applicability of TSI model in real-world scenarios.
Bibliographical notePublisher Copyright:
© 2020 Elsevier Inc.
Copyright 2020 Elsevier B.V., All rights reserved.
- Anomaly detection
- Data analysis
- One class support vector machine
- Smart meter
ASJC Scopus subject areas
- Control and Systems Engineering
- Theoretical Computer Science
- Computer Science Applications
- Information Systems and Management
- Artificial Intelligence