| With the information resources on the network released, how to properly and quickly accessthe resources needed to become an important issue. To retrieve the drugs’ information, the mostof the retrieval engine use the the key-words matching method to carry out the the retrieval ofdocumentation and resources. However, simply using keyword matching method ignores thekeyword’s synonyms and related words matching, leading to some of the resources can not beretrieved. In addition, the keyword matching method can not correctly understand the the syntaxof the structure and semantic information of the sentence, leading to a large number of resourcesthat are inconsistent with the user demand is retrieved, even search out the resources of thedeviation from the user’s retrieval intention.Response to these problems, this paper proposes semantic understanding of the searchstatement in the rational use of medicines query semantic analysis on the search statement,including the retrieval statement carried natural language processing, such as segmentationprocessing, lexical tagging, syntactic dependence analysis, the semantic analysis the result of theconversion into the corresponding (subject, predicate, object) in the form of triplet sequences,and triple converted into a corresponding sequence in accordance with the specification of theSPARQL query SPARQL query, in order to improve the retrieval recall and precision rates.The work completed are listed:(1)Analysis the document of Chinese drugs’ name which had been translated by our teamfrom Drugbank document and extract the name of the drugs, protein names and other medicalvocabulary from the document;(2)Extract part of the Drugbank data to construct the corresponding node knowledge baseand the knowledge base of rules mapping keywords to RDF nodes, and the nodes’ knowledge areclassified;(3)Extract semantic information of natural phrases, and convert the natural sentence intoSPARQL queries basing on responding knowledge base and SPARQL specification. |
PDF Full Text Request