【摘要】 基于概念类别属性,在Protege平台下构建了茶领域本体,并实现基于茶领域本体的DocOnto文本分类器。在该分类器上对茶文档、酒文档和比萨文档进行分类实验,并与朴素贝叶斯分类器的实验结果对比,表明DocOnto分类器在综合查准率相当的情况下,有效地提高召回率,获得更高的F1指标。更多还原

【Abstract】 Tea domain ontology was constructed and an ontology-based text classifier named DocOnto was implemented based on classes of concept in Protege.In the experiment,tea texts,wine texts and pizza texts were respectively classified to their corresponding categories by DocOnto,and the experimental results comparisons were also made between DocOnto and Naive Bayes.It shows that at the equivalent level of comprehensive precision,DocOnto can improve the recall effectively and get the higher F1 index.更多还原