Abstract
We study frequent pattern mining from positional data streams. Existing approaches require discretised data to identify atomic events and are not applicable in our continuous setting. We propose an efficient trajectory-based preprocessing to identify similar movements and a distributed pattern mining algorithm to identify frequent trajectories. We empirically evaluate all parts of the processing pipeline.
| Original language | English |
|---|---|
| Title of host publication | New Frontiers in Mining Complex Patterns |
| Number of pages | 15 |
| Publisher | Springer Verlag |
| Publication date | 2015 |
| Pages | 102-116 |
| ISBN (Print) | 978-3-319-17875-2 |
| ISBN (Electronic) | 978-3-319-17876-9 |
| DOIs | |
| Publication status | Published - 2015 |
| Externally published | Yes |
| Event | 3rd International Workshop on New Frontiers in Mining Complex Patterns - NFMCP 2014 - Nancy, France Duration: 19.09.2014 → 19.09.2014 Conference number: 3 |
Research areas and keywords
- Informatics
- Pattern Mining
- Dynamic Time Warping
- Positional Data
- Frequent Episode
- Event Stream
- Business informatics
Fingerprint
Dive into the research topics of 'Mining positional data streams'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver