The Research Of Semantic Retrieval Based On Domain Ontology

With the rapid development of Internet and the quickly growth of network information, traditional semantic system can’t understand semantic, either retrieving lots of irrelevant information or missing much information. It hasn’t retrieved the resources which users need quickly and exactly from mass of internet resources.Semantic retrieval system based on the domain ontology can understand semantic of document and user’s query intention, so it is applied widely. But it also has some disadvantages, for example, borer、 fertilization and irrigation, they are concepts of rice field and also are the concepts of corn field at the same time, so they have some field’s semantic, when query these vocabularies, their precision are low. The article aims at studying semantic retrieval based on domain ontology, main works are following:1.The construction of the ontology of the domain. The article takes rice as example, adopting seven footwork domain ontology modeling which is developed by Stanford University to construct rice domain ontology.2.The works of semantic annotation. Divided the concept vocabularies of domain ontology into two kinds, the first is unique concept which are possessed by the only field; the second is ordinary concept which can appear in some fields. After using annotation algorithm based on domain ontology to calculate the key’s weight, taking unique concept as guidance to calculate ordinary concept’s weight again.3.Making query extension to user input word by the domain ontology, then forming query vector. Query extension can be divided to two cases:(1) if query term is match to concept of ontology, then combining with ontology to extend semantic;(2) if query term isn’t match to concept of ontology, we can search the concept of ontology from documents to extend which appear at the same time to query term and have high frequency.The article constructs and realizes the information retrieval system based on the rice domain ontology. As is shown in experiments, the system in guarantee recall rate basis improves precision rate of ordinary concept.