Font Size: a A A

Ontology System Based On Medical Events

Posted on:2015-01-27Degree:MasterType:Thesis
Country:ChinaCandidate:J SunFull Text:PDF
GTID:2298330431995857Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Currently there are a large number of medical literatures on the Internet, such asdrug instructions and disease treatment guidelines, etc. The event data in thesemedical literatures has great significance in medical treatment, clinical monitoringand some other medical field. However, given the vast amounts of medical literature,it is clearly unrealistic to turn these data into structured data that can handled by anymachine manually.This paper proposed a solution of ontology system based on medical eventsthrough the deeply research on Ontology building, which includes several aspectssuch as data conversion, storage, query, etc. The main purpose of it is to translatenatural language data into structured data which is capable for computer to handle,and then these structured data could be used by the upper application. The mainresearch program is divided into the following steps:1) Extract the useful information from medical guide data with segmentationsystem which is developed by other members of this team.2) Summed the medical event model, using the data model and relational-object database PostgreSQL to build a storage system to store the informationextracted in front.3) Analyzing the characteristics of several SPARQL matching graph pattern tocomplete the mapping of SPARQL to SQL, which allows users to submitSPARQL queries to database and returns the corresponding result set.4) Establish a data export module, which could export information fromdatabase into textual information. When it comes to the text, the system useNamed Graph to describe medical events, data with this structure is effectiveto reduce data redundancy and improve the efficiency of the query. On theother hand, Named Graph could be used to describe the relationship betweendifferent events, and that makes different information could establish contactwith each other.Finally, this paper valued the system by compare it with Sesame(a ontologyquery system), and get a better search result and a more streamlined data set, whichvalidate the feasibility of this method in further.
Keywords/Search Tags:Semantic Web, Ontology building, Event Model, SPARQL graphpatterns, PostgreSQL
PDF Full Text Request
Related items