This repository contains multiple algorithms and the experiments studied and presented in our work "Spatiotemporal Group Anomaly Detection via Graph Total Variation on Tensors".
- HoRPCA:
src/algos/horpca_singleton.py - LR-STSS:
src/algos/lr_stss_modified.py - LR-STS:
src/algos/lr_sts.py - LR-SSS:
src/algos/lr_sss.py
Experiment setup scripts, control and independent variables and hyper-parameters of the experiments can be found in the following files:
- Sythetic Experiment 1:
testbench/exp1_number_of_anomalies_hpcc.ipynb - Sythetic Experiment 2:
testbench/exp2_anomaly_duration.ipynb - Sythetic Experiment 2:
testbench/exp3_anomaly_radius.ipynb - NYC Yellow Taxi Records 2018 Anomaly Detection"
testbench/nyc_optuna_hp_search.ipynb
M. Indibi and S. Aviyente, "Spatiotemporal Group Anomaly Detection via Graph Total Variation on Tensors," ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Seoul, Korea, Republic of, 2024, pp. 7035-7039, doi: 10.1109/ICASSP48485.2024.10448399.
Mert Indibi