Research On Temporal Reasoning Algorithm Based On Temporal Ontology

Time is an important dimension to describe the process of information change.Most information in the real world has temporal characteristics.The analyzing and mining of effective information in temporal information has become a research hotspot in the field of artificial intelligence.With the improvement of users' information quality requirements,information semantics has become the main mode and inevitable trend of information management,utilization and sharing.Ontology is an effective way of information management,sharing and knowledge modeling.However,static ontology does not deal with time in detail and cannot express implicit time.In reality,a large amount of temporal information is not fully utilized.Quickly obtaining implicit information from temporal information is an indispensable part of practical applications.Therefore,how to use such information efficiently to infer ontology knowledge is a problem that needs to be solved.In this paper,temporal data representation based on temporal ontology is used for temporal reasoning.The time ontology realizes the formal description of the sharing concept and its relationship in the time domain,strengthens the common understanding of domain knowledge,and has good reasoning ability.Time and events are inseparable.The temporal ontology obtained by fusing the time ontology and the event ontology can better represent the change process of things.In order to obtain more effective implicit relationships,it is not enough to rely on temporal ontology.Rules can enhance the inference effect of ontology,and use the rule inference engine to make temporal inferences on fact data and rules.Therefore,from the semantic point of view,this paper focuses on the reasoning of time information,researches on temporal data representation modeling,temporal reasoning algorithms,etc.,and has made the following progress in theoretical analysis of key problems and implementation of technical research:(1)Modeling temporal data representation methods.By using the upper body of BFO and OWL-Time,a general tense ontology model is constructed,and the corresponding rules are constructed.The temporal data representation method based on temporal ontology is realized.(2)Improve the ReteOO algorithm.Although the ReteOO algorithm has been successfully applied to the Drools rule inference engine,the algorithm still has room for improvement in reasoning efficiency.According to the principle of locality,in the rule network constructed by ReteOO,the inference efficiency is improved by introducing a fast memory area between the root node and the type node,and establishing an index queue in the fast memory area.(3)Design and implement a temporal inference prototype system based on temporal ontology.Taking the human resources field as the application background,the temporal data representation method in the field and the improved ReteOO are used for experimental analysis.Finally,the availability and effectiveness of the proposed system framework are verified by experiments,and a reasonable evaluation of the system is made through comparative experiments.