Research On Quantitative Reasoning Method For Event Ontology

Due to the knowledge represented by event ontology is closer to human's understanding and analysis about knowledge,event ontology gets more and more attention from academia.Meanwhile,time and space are fundamental elements of event,and semantics of event can be expressed more exactly with them.So,event representation and reasoning based on time and space become hotspots of research about event ontology in recent years.In the event reasoning,relation reasoning and knowledge mining are restrained by the weak reasoning of ontology.It makes more and more scholars user inference rules in event mining.However,current rule-based reasoning ways are not involved in precise quantitative information.In other words,the existing rules can only reason out hidden relationships according to known relations,and they cannot determinate the certain relationship with properties of entity or object.Definition of event is extended with time and space in this paper.Both qualitative reasoning and quantitative reasoning are involved by defining temporal relationship and spatial relationship quantitatively.In addition,the rules of relation definition and combination of reasoning are summarized in this research.Meanwhile,rules are merged and simplified with the transitivity,symmetry and inversion of event relationship.At last,30 quantitative reasoning rules and 72 combinatorial reasoning rules are realized with math functions in SPARQL.The experiment result shows that,the quantitative reasoning method can effectively dig relationships between entities in event ontology.Meanwhile,the effectiveness and performability are verified with this experiment.