基于规则与统计的本体概念自动获取方法研究

Research on AUtomatic Ontological Concept Extraction Based on Rules and Statistics

为获取中文领域本体的概念提出了基于规则匹配和统计方法相结合的学习模型，充分利用现有的自然语言处理技术和统计学习方法，从领域文本中通过语义串切分、规则匹配、领域归属度分析和概念约简算法自动获取领域概念。该方法解决了现有中文本体学习方法对领域词典的依赖以及无法获得短语式特定领域概念的问题，同时解决了领域概念筛选问题。实验证明了该方法的有效性。

A methodology based on rules and statistics is proposed to acquire Chinese ontological concepts. This methodology uses NLP techniques and statistical algorithms to extract domain-specific concepts from texts through semantic string segmentation, rules matching, analysis of domain consensus and relevance, and concept reduction. This methodology resolves current problems in Chinese ontology learning, such as dependency on domain lexicons, difficulties for phrasal concepts acquisition, and domain-specific concepts filtering, which can be demonstrated by experimental results.