Relation Extraction And Retrieval Based On Ontology

With the rapid development of the computer technology and network technology, online information increases exponentially with bandwidth. In such a society of explosive information, people want to get much information through the network, and in this context relation extraction was born and developed. Generally speaking, a customer inquires information through the relation extraction. Relation extraction can offer the relation between entities and relevant sorting web according to a traditional method, although sometimes it can not meet user queries. In order to provide satisfactory service for customers and inquire more implicit semantic relations, in this paper we mainly research a new method for relation extraction and retrieval which based on ontology. It can offer intelligent service, such as the extraction of more implicit semantic relations and more satisfactory retrieval results to customers.It is found that there are a large number of implicit semantic relations can not be extracted through the study of traditional relation extraction technology. The traditional relation extraction gets the relation base and extracts information directly through the method such as patern matching or machine learning. In other words, traditional relation extraction just analyzes syntax information and no semantic kownledge, missing a large number of valuable information. To address this serious problem, this paper puts forward relation extraction and retrieval based on ontology. In order to adapt the ontology to relation extraction, firstly we establish a relation ontology in this paper. For the sake of it not only rich in semantic information but also based on relation concept, the expanded relation base built by semantic reasoning performs better, achieving a seamless integration to relation extraction. On the other side, in the process of relation retrieval, we offer SSR, an effective way to measure the semantic similarity and relativity, and the ontology-based entity expansion, which can improve the precision and recall rate. In terms of ranking, traditional method only takes account of objective factors, which could not satisfy the customers practically. For the reason, this paper comes up with a Semanteme Preferred Ranking (SPR) method based on the customer’s query demands, which considers the realtion types, web pages, and also the semantic matching degree to user’s demand. Thus SPR makes the system understand customers’information demand more accurately and improve the quality of user experience (UE), finally meets the uers’ query demand.According to the experiment’s results, relation extraction based on ontology achieves a comparative satisfactory effect. The retrieval and ranking of relation incorporated with semantic information is superior to traditional method and meets the needs of the users for searching information.