本体是领域知识的有效组织和描述，本体学习则是实现本体自动构建的方法体系和技术集合。本文以本体学习理论为指导，提出了一种以文档-术语空间为核心、形式概念分析（FCA)为手段的中文领域本体层次结构自动构建的有效方法，并以“白血病”领域为例，对面向学科资源的医学专业术语层次关联的抽取进行了详细论证，具体包括专业术语的抽取和筛选，术语文档关联的修正等数据清洗过程；文档术语矩阵的建立，领域概念格的自动生成，以及概念格中术语属性的层次关联建立等FCA过程；术语层次关联的自动OWL描述和存储，和领域本体的概念检索和可视化展示过程等。%O ntology is theeffectiveorgan izatio n and descrip tio n for domain know ledgeand O ntology Learnin g ( O L ) is themethodology and tech n olo gy to con struct O ntology a uto ma ticaly. WiththeO L theo ry as a g uid e,this pap er proposes an effectivemethod ,which is withdocumen ts-terms spaceas a coreand withF o rma l C oncept A n a lysis ( FCA) as a mea n s,to con struct h ierarchy struc tureofC hineseD oma in O ntology a uto ma ticaly. Taking “ leuk emia ” field for an example,it in d eta il demonstrates theextra ctin g process on h ierarchy relation ship of Med ic a l professio n al terms o rien tedd is ciplin es res o urc e,which s p ec ific a l y contains 3 processes. F irs t is thedata c lea rin g process as in itia liz a tio n in c lud in g extra ctin g and filterin g of professio n al terms ,and amendment ofassociation ofterms fromdocuments. Second is theFCAprocess in c lud in g b uild in g ofdocumen ts-terms ma trix , a uto matic generation ofdomain concept la ttic eand con structio n ofh iera rch y relation s h ip ofp rop erties fro mterms in concept la ttic e. Third is theterms ontology d escrip tio n process in c lud in g automatic OWL descrip tio n and storageofh ierarchy associations ofterms ,concep t searching and vis ualy display in g ofdomain ontology.
展开▼
