The algorithm proposed in the paper 'Semantic Discord: Finding Unusual Local Patterns for Time Series' is up to 3 orders of magnitude faster than a brute force algorithm on real world data.
Sources
- Semantic Discord: Finding Unusual Local Patterns for Time Series (arxiv.org)
- https://mason.gmu.edu/~lzhang18/papers/SDM20_anomaly.pdf (mason.gmu.edu)
- https://www.semanticscholar.org/paper/Semantic-Discord:-Finding-Unusual-Local-Patterns-Zhang-Gao/ed99a343557cb9e2578d824c22fb0f15facd03c0 (semanticscholar.org)
- https://dx.doi.org/10.1137/1.9781611976236.16 (dx.doi.org)