Name
Title | ||
|---|---|---|
Smooth and crispy: integrating continuous event proximity calculation and discrete event detection. |
Abstract | ||
|---|---|---|
Using Complex Event Processing for event detection is appealing for emerging applications in home care and health care because hazards and anomalies need to be detected to alarm a care giver as fast as possible. It is important for a care giver to receive alarms if anomalies and hazards are detected. However, false positives and false negatives can imply high costs and it is important to personalize the application for every individual instantiation. We introduce in this paper the concept of Event Proximity (EProx) to address these challenges. EProx quantifies the closeness of the current situation to an event of interest. EProx enables the system and its users to react to anomalies and hazards before they actually happen. We present a template for EProx calculation and exploit a home care use case to explain how the template is instantiated in Esper. We use synthetic activity traces of a care receiver to verify and demonstrate the usefulness of event proximity calculation. |
Year | DOI | Venue |
|---|---|---|
2016 | 10.1145/2933267.2933302 | DEBS |
Field | DocType | Citations |
Health care,Data mining,ALARM,Closeness,Computer science,Complex event processing,Real-time computing,Exploit,False positives and false negatives,Care giver | Conference | 0 |
PageRank | References | Authors |
0.34 | 5 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Stein Kristiansen | 1 | 72 | 11.86 |
| Thomas Plagemann | 2 | 496 | 69.84 |
| Vera Goebel | 3 | 327 | 47.86 |