论文部分内容阅读
针对传统矢量空间模型文本特征值的计算,给出将文本的评价由基于语法的词条空间转化为概念空间的方法和策略:基于领域本体,依据本体概念间的各种关联,先以一定的映射规则,将词条映射到领域的概念术语空间;然后用概念统计和语义归纳替代传统的词频统计,从概念语义的层次计算文本的特征矢量。实验证明,基于本体概念的矢量检索模型能够有效地表达文本的语义内容,获得更好的检索效果。
Aiming at the calculation of textual eigenvalues of traditional vector space models, a method and strategy of translating textual evaluations from grammar-based lexical spaces into conceptual spaces are given: based on domain ontologies, based on the various connections between ontology concepts, Mapping rules, the entries are mapped to the domain of the conceptual space of terms; and then use the concept of statistics and semantic induction to replace the traditional word frequency statistics, the level of conceptual semantic computing text feature vector. Experiments show that the vector retrieval model based on ontology concept can effectively express the semantic content of texts and obtain better retrieval results.