Research On The Reasoner Of Intelligent Question Answering System

With the rapid development of the Semantic Web, ontology-based applicationscome out increasingly. Semantic Web makes data on the Web possess machinereadable semantic to facilitate the human machine interaction. Ontology reasoner isone of the key supporting technologies to implement the Semantic Web, but due to thesingle ontology reasoner lacks reasoning precision and adaptability, this paper putsforward the concept of the ontology multi-layer reasoner, designs multi-layer reasonermodel, researches multi-layer reasoner implementation mechanism, uses the Pelletreasoner and Jena custom rule, and implements multi-layer reasoner semanticreasoning. As a ontology querying and reasoning engine, ontology multi-layerreasoner has two clear tasks: one is to be responsible for executing the ontologyconsistency checking and classification reasoning function and so on, then using Jenacustom reasoner to perform uncertainty reasoning; and the other one is to accept userontology query command, execute the query and submit the query results to the user.Experimental results show that this method can make up for the deficiency caused bya single semantic reasoning techniques used in traditional reasoner, and is moreeffective than a single reasoner in digging out the hidden information of the ontology.The work of this paper mainly focuses on the following aspects:First of all, the article introduces ontology, description language and ontologyquery language, and then comprehensively and systematically analyses the currentontology reasoner systems, covering the comparison of ontology reasoner’sfunctionality, category and typical reasoner, and ontology reasoner implementationtechnology and so on. Besides, it summarizes the latest tren ds of ontologyreasoner-these can provide a useful reference for ontology reasoner application anddeveloper user.Second, based on systematically researching the reasoner mechanism of genericreasoner Jena and specialized reasoner Pellet, Racer, Fact++, d esigns ontologymulti-layer reasoner architecture model, writes its semantic reasoning step, it not onlycan resolve and execute the ontology query command, but also be able to performontology reasoning functions of all kinds of ontology description langu age. Theauthor combines some mainstream ontology reasoning functions of ontology reasoner,presents the set of multi-layer reasoner API scheme, and provides a unified interface for the interaction between the ontology reasoner and application.Then, based on multi-layer reasoner designs and implements the questionanswering system, for the first time multiple reasoners are applied to the questionanswering system. And it suggests a set of more comprehensive experimental programfor ontology reasoner system test comparison, from the aspect of function, user, anddevelopment. The experiment proved that the testing program is feasible and effective.Finally, summarizes the main research and experiments above, points out the directionof the further research, and illustrates this article has some theoretical and practicalsignificance for the Semantic Web reasoner research and development.