A SKOS-based Multi-granularity Semantic Annotation Method For Academic Text

The traditional knowledge organization system can't provide the semantic information which can be understood by the computer,this problem restricts the organization,collaboration and sharing of knowledge on the Internet.In order to solve this problem,Tim Bemers-Lee proposed a semantic network to transform information processing from simple grammar processing to complex semantic processing.With the development of semantic web,semantic annotation has become more important.At present,ontology-based semantic annotation method has achieved remarkable accomplishments.However,in the application we found that to embody all the concepts in one ontology requires a lot of work,especially has to use the knowledge of many experts in the certain filed.This becomes a barricade to wildly applicate the ontology-based semantic annotation.Fortunately,W3C has published the Simple Knowledge Organization System(SKOS)to provide a way to solve this problem.In addition,the common annotation methods mostly use coarse-granularity,large amount of internal information contained in the resources can't be retrieved,filtered and extracted.In order to solve this problem there is another kind of annotation methods use fine-granularity.This kind of methods can indicate the minimum unit of the resources and feedback to the users the most direct knowledge.Whatever the coarse-granularity or fine-granularity annotation methods can only achieve a single granularity knowledge organization model,it hard to meet the users' needs of multi-granularity.we can redesign the process to achieve the multi-granularity semantic annotation.Firstly,this paper summarizes the theoretical foundation of semantic annotation and SKOS,including the granularity organization of semantic annotation,the process,technology and tool of ontology-based semantic annotation,and SKOS description of controlled vocabulary.These studies will all be important reference for this paper.Secondly,based on the related research,this paper analyzes the four key issues of the SKOS-based multi-granularity semantic annotation method for academic text,and then propose its franework and description.Thirdly,this paper presents the detail process of SKOS-based multi-granularity semantic annotation method for academic text,including the granularity instruction of academic text resources,the multi-granularity segmentation,the selection of multi-granularity annotation words and SKOS description of multi-granularity annotation result.Finally,this paper selects the appropriate academic text resources as objects,sets up the experiment group and the control group respectively,uses the SKOS-based multi-granularity semantic annotation method to carry on the empirical study.At the end evaluates the annotation results to verify the annotation effect of this method.