DISCOVERY OF MULTI-DOMAIN ANOMALOUS TEMPORAL ASSOCIATIONS.
Loading...
Links to Files
Permanent Link
Collections
Author/Creator
Author/Creator ORCID
Date
2017-01-01
Department
Information Systems
Program
Information Systems
Citation of Original Publication
Rights
This item may be protected under Title 17 of the U.S. Copyright Law. It is made available by UMBC for non-commercial research and education. For permission to publish or reproduce, please see http://aok.lib.umbc.edu/specoll/repro.php or contact Special Collections at speccoll(at)umbc.edu
Distribution Rights granted to UMBC by the author.
Distribution Rights granted to UMBC by the author.
Abstract
Temporal data can capture the behavior of phenomena such as accidents along a highway, weather trend such as precipitation or snow totals in a region over time. Traditional temporal data mining has looked at patterns, such as anomalies, in each temporal data stream. However, to study real world phenomena and interrelationships between them, in this theses, we propose a novel approach to discover the temporal relations between multiple distinct domains represented by multiple distinct temporal data collected at a location. Our goal is to discover the relationship between distinct domains using interesting temporal events in them. These interesting temporal events are mined using traditional temporal anomaly detection methods. Relations between two application domains are not always simple since there can be some time-delay in these relationships. So, focusing on relations found using intersecting time events alone is not sufficient. Hence, we employ the concept of not only direct overlap but also proximity between temporal events across domains to find the direct and time-delayed relationships. We have achieved an optimistic result after our experiment on MATCH, NJDOT and weather data.