Research Article
dSalmon: High-Speed Anomaly Detection for Evolving Multivariate Data Streams
@INPROCEEDINGS{10.1007/978-3-031-48885-6_10, author={Alexander Hartl and F\^{e}lix Iglesias and Tanja Zseby}, title={dSalmon: High-Speed Anomaly Detection for Evolving Multivariate Data Streams}, proceedings={Performance Evaluation Methodologies and Tools. 16th EAI International Conference, VALUETOOLS 2023, Crete, Greece, September 6--7, 2023, Proceedings}, proceedings_a={VALUETOOLS}, year={2024}, month={1}, keywords={Outlier detection Data streams Unsupervised learning Python C++}, doi={10.1007/978-3-031-48885-6_10} }- Alexander Hartl
Félix Iglesias
Tanja Zseby
Year: 2024
dSalmon: High-Speed Anomaly Detection for Evolving Multivariate Data Streams
VALUETOOLS
Springer
DOI: 10.1007/978-3-031-48885-6_10
Abstract
We introduce dSalmon, a highly efficient framework for outlier detection on streaming data. dSalmon can be used with both Python and C++, meeting the requirements of modern data science research. It provides an intuitive interface and has almost no package dependencies. dSalmon implements main stream outlier detection approaches from literature. By using pure C++ in its core and making the most of available parallelism, data is analyzed with superior processing speed.
We describe design decisions and outline the software architecture of dSalmon. Additionally, we perform thorough evaluations on benchmarking datasets to measure execution time, memory requirements and energy consumption when performing outlier detection. Experiments show that dSalmon requires substantially less resources and in most cases is able to process datasets between one and three orders of magnitude faster than established Python implementations.
