Sebastian Schmidl

Results 62 issues of Sebastian Schmidl

Add Series2Graph++ algorithm to this repository. - Source code: https://github.com/HPI-Information-Systems/S2Gpp Either build pip-Package for S2G++, publish it to PyPI, and just use this package in an algorithm folder; or add...

:medal_sports: medium
comp: algorithms

Currently, we install R packages using the default `install.packages`. This gives non-reproducable results because the packages get updated frequently. We should use `renv` or `packrat` to manage R package versions....

:bug: bug(-fix)
:medal_sports: low
comp: :package: CI / project / packaging

Add AnomalyTrans as a supported algorithm: Paper: https://arxiv.org/pdf/2110.02642.pdf (accepted for ICLR 2022) Code: https://github.com/thuml/Anomaly-Transformer

enhancement
help wanted
comp: algorithms

Adds the GDN algorithm from https://github.com/d-ailin/GDN (Paper: https://doi.org/10.1609/aaai.v35i5.16523) Fixes #15

:medal_sports: medium
comp: algorithms

Add the newly published DCDetector as supported algorithm: - Paper: https://arxiv.org/abs/2306.10347 (accepted for SIGKDD 2023 in August) - Code: https://github.com/DAMO-DI-ML/KDD2023-DCdetector - DOI: https://doi.org/10.1145/3580305.3599295

enhancement
help wanted
:medal_sports: medium
comp: algorithms

Audibert et al. proposed a new DL-based anomaly detection algorithm for multivariate time series in https://doi.org/10.1145/3394486.3403392. We should include this algorithm in the repository. A community implementation can be found...

enhancement
help wanted
:medal_sports: medium
comp: algorithms

There is a new promising multivariate anomaly detection algorithm called GDN. We should add its implementation to this repository. - Paper: https://doi.org/10.1609/aaai.v35i5.16523 - Implementation: https://github.com/d-ailin/GDN --- Thanks to @2er0 for...

enhancement
help wanted
:medal_sports: medium
comp: algorithms

Code: https://github.com/DAMO-DI-ML/CIKM22-TFAD Paper: https://doi.org/10.1145/3511808.3557470

enhancement
help wanted
:medal_sports: medium
comp: algorithms

Use an existing time series as base oscillation for a new dataset. This allows injecting our anomalies into existing time series. The file must be formatted in our canonical file...

enhancement
:medal_sports: very low

Take a subsequence and flip it vertically: ![image](https://user-images.githubusercontent.com/10573700/217221460-ce98f711-d5ec-4c90-b253-a94f27babfda.png) This anomaly can be used for periodic datasets only. Add an option for a smooth transition if the start and end does...

enhancement
good first issue
:medal_sports: low