The QuAD system proposed in the paper detects anomalies in time series based on representations learned from time series prediction. The core of the system is a time series prediction network, QuADNet. This repository contributes to the ease of use of QuADNet.
We recommend that you check src/run.py first.
The file is trained and tested by QuADNet based on the variables received by the parser.
The list of variables received by parser can be viewed by running python run.py -h or by checking the src/parse.py file.
Data should be prepared in data_dir under the names train_data.npy and test_data.npy.
The train_data.npy file is loaded during training, and the test_data.npy file is loaded during testing.
@INPROCEEDINGS{iecon-quad-2022,
author={Ishizone, Tsuyoshi and Higuchi, Tomoyuki and Okusa, Kosuke and Nakamura, Kazuyuki},
booktitle={IECON 2022 – 48th Annual Conference of the IEEE Industrial Electronics Society},
title={An Online System of Detecting Anomalies and Estimating Cycle Times for Production Lines},
year={2022},
volume={},
number={},
pages={1-6},
keywords={Training;Meters;Energy consumption;Power demand;Neural networks;Estimation;Benchmark testing;anomaly detection;key production performance indicator;quasi-periodicity;attention mechanism;smart meter},
doi={10.1109/IECON49645.2022.9969061}}