Abstract

Techniques to compute event class weights are crucial to aid and encourage the re-use of publicly available event ontology. This paper presents a method of event class weights for event ontology in terms of the characteristics of event ontology, which considers not only the associative types between event classes, but the event class elements such as objects, locations and event triggers. The experimental results show that the average precision is 0.78 and the average deviation is 1.212.

Paper Details