论文部分内容阅读
为获取领域本体并量化概念关系的可信度,提出了一种基于W eb挖掘的学习模型。通过可扩展的模式集和分布语义模型获取本体主干,使用关联规则发现概念间的一般关系,对候选本体进行修剪和合并。模式可信度、概念语义距离与关联特征决定了概念间关系的可信度。通过“文本分析本体获取文本扩充”的迭代过程,优化模型参数和阈值。该模型解决了现有本体学习方法对词典或核心本体的依赖性、以及不能对关系进行可信度量化的问题。实验证明了所提出模型的有效性。
In order to obtain the domain ontology and quantify the credibility of the conceptual relationship, a learning model based on the W eb mining is proposed. Ontology stems are obtained through scalable pattern sets and distributed semantic models, association rules are used to discover general relationships between concepts, and pruning and merging candidate ontologies. The model credibility, the concept of semantic distance and the correlation features determine the credibility of the relationship between concepts. Optimize model parameters and thresholds through an iterative process of “textual expansion ontology for text analysis.” The model solves the dependence of existing ontology learning methods on dictionaries or ontologies, and the problem of not being able to quantify the trustworthiness of relationships. The experiment proves the validity of the proposed model.