Application Research Of Semantic Retrieval Technology In Exploration And Production Portal

As the amount of data in Exploration and Production Portals increases year by year,the keyword-based portal information retrieval system currently used can no longer meet users requirements in terms of query response speed and precision and recall rate.This dissertation proposes an ontology-based semantic information retrieval method,which improves the shortcomings of current retrieval modules in portals.This dissertation analyzes the advantages of ontology-based intelligent semantic retrieval technology in the application of enterprise portal,and on the basis of understanding the demand of the Exploration and Production Portal for semantic retrieval technology and exploring the key technologies of semantic retrieval,the dissertation puts forward the idea and the overall design scheme of the semantic retrieval system for the Exploration and Production Portal.Firstly,this dissertation combines web crawler and Chinese word segmentation technology to process the information of portal pages,uses core package of Lucene for indexing and extracts keywords in TF-IDF model.Then,the keywords and user input terms are compared with the ontology model,and the semantic relevance of different instances is inferred by the concept and the relationship between concepts.Finally,the semantic analysis and the word-frequency position weighted sorting algorithm are combined to calculate the relevance of the web pages and the search terms more objectively and accurately.This dissertation applies the ontology to the PetroChina Exploration and Production Portal,designs and develops a semantic retrieval system for Exploration and Production Portals.With the system,users can view the content details in the search list and the semantic relevance of the search list and query terms.The experimental results show that the semantic similarity after the expansion of query terms is improved to 21.2%,and the precision rate and the recall rate of user searches is increased to 77% and 68%.Enterprises and users can locate massive information quickly and accurately,and query response speed is improved significantly.This will better reflect the level of informatization services for Exploration and Production Portals.