This dataset was created by sifting through the REFIT dataset to detect load anomalies; the rules for labelling anomalies are described in the accompanying ICASSP'19 paper, which should be referenced if the dataset is used. Five of the 20 houses of the REFIT dataset were included in this dataset, as they contained the largest number of detected anomalies. These are Houses 1, 10, 16, 18 and 20. At the time of release, this is the first detailed annotated dataset of anomalies within publicly available electrical load measurements. These are real anomalies, not simulated ones and are extremely useful in understanding anomalous behaviour of electrical appliances, as measured by smart meters. Anomalous behaviour of the following appliances is included in this dataset: Refrigerator, freezer, fridge-freezer, dishwasher, washing machine, tumble dryer, electrical heater and microwave. When using this dataset, please cite the following paper: H. Rashid, V. Stankovic, L. Stankovic and P. Singh, "Evaluation of Non-Intrusive Load Monitoring Algorithms for Appliance-level Anomaly Detection," Proc. IEEE 44th Int. Conf. on Acoustics, Speech and Signal Processing (ICASSP), Brighton, UK, May 2019.
Data will be made available on 12/05/19 after the conference has closed.
|Date made available||6 Feb 2019|
|Publisher||University of Strathclyde|
|Temporal coverage||Oct 2013 - Jun 2015|
|Date of data production||2017 - 2018|