Font Size: a A A

Research On Construct Of Domain Ontology Based On Hierarchical Demand And Method Of Semantic Annotation

Posted on:2011-07-17Degree:MasterType:Thesis
Country:ChinaCandidate:Z ZhangFull Text:PDF
GTID:2178330332982024Subject:Business management
Abstract/Summary:PDF Full Text Request
In the network information age, the origin of information relies more on the Internet which is an information resource warehouse.The information resource that constituted by trillion homepage set a higher request to the capacity of information processing.That the natural language can only be understand by the human being becomes the bottleneck of information development,it also caused the contradiction between information inflation and the user information insufficient.The semantic web is the expansion and extends of World Wide Web, and it is also an important technology for automation information processing.The widespread application of semantic web relies on both development in method domain ontology constructing and the semantic annotation technology.Ontology is a kind of modelling tools which can describe the conceptual model of information system on the semantic level,which can provide common understanding to the domain knowledge, build domanial frame system of concept and express it in the formalized way.It have important application in the area of construction of intelligent retrieval system and semantics web, knowledge engineering and knowledge management.At present, following the apply of domain ontology deeper and deeper many construction of ontology have presented,however the method still exists some flaws,for instance the lack of standard and Engineering ideology,non-standard etc.Semantic annotation is the process increasing the standardized knowledge to the information under the instruction of ontology,using defined concept and the individual to mark the documental information, then understanding the documents.Generally,the present method of semantic annotation based on the rule way and based on the natural language way, which cannot meet the requirement of large-scale semantic annotation.Based on the analysis of characteristic of domain lamination, this article proposed a demand analysis method based on the domain lamination that breach the limitation of specific domain.It is carried on according to the decomposition of domain's level, which make up deficiency in former methods of domain ontology, that provide one kind of mentality to achieve Engineering in constructing domain ontology.On the basic of the demand analysis method of domain lamination, this article proposed the method based on the domain level demand analysis,which profits from the existing domain ontology constructing method and based on the software engineering theory.This method divides the domain construction process into four stages,that is demand analysis, the ontology designing, the ontology coding, the examination and appraisal.Each stage has the corresponding report and documents, and easy to maintain and the expansion, this meets the project development requirement and suits in the multi-person's cooperation development. Using the method that we have proposed, this article uses protege to create the energy economy domain ontology, which confirms our method's feasibility and validity.Further, this article summarizes and divides current semantic annotation methods from two dimensions which are the automaticity way and the semantic annotation way. Then, this article presents an automatic semantic annotation method that based on the domain ontology, which divides semantic annotation into two stages:entity recognition stage and relational extraction stage. The previous stage uses the ambiguous word eliminating method to eliminate the emergence that the same concept have different meaning in different domain ontology; The latter stage has used method of semantic relations extraction that is based on the domain ontology, which synthetically utility the triple of structure in domain ontology and the type of attribute value. And this article verify the validity of our method through experiment.
Keywords/Search Tags:Domain Ontology, Semantics Annotation, Relation Extraction, Creation Method
PDF Full Text Request
Related items