Structuring ordered nominal data for event sequence discovery
Conference Paper
Overview
Identity
Additional Document Info
Other
View All
Overview
abstract
This work investigates using n-gram processing and a temporal relation encoding to providing relational information about events extracted from media streams. The event information is temporal and nominal in nature being categorized by a descriptive label or symbolic means and can be difficult to relationally compare and give ranking metrics. Given a parsed sequence of events, relational information pertinent to comparison between events can be obtained through the application of n-grams techniques borrowed from speech processing and temporal relation logic. The procedure is discussed along with results computed using a representative data set characterized by nominal event data. 2010 ACM.
name of conference
Proceedings of the 18th ACM international conference on Multimedia