基于本体学习的标签推荐方法研究

Method Research on Tag Recommendation Based on Ontology Learning

由于标注过程简单,Web上标注系统的使用逐渐增长,但是,随意定义的标签缺少标准并且语义模糊.为改善标签系统推荐效果,帮助用户组织、管理及分享网络资源,提高检索效果.提出基于用户标注信息的本体学习方法,针对不同映射情况,设计对应的本体学习模型和语义歧义消除模型,通过基于本体表示标签的语义信息和基于扩展本体语义关系的标签排序方法推荐标签.实验证明,召回率和精度都有提高,方法具有较好的可行性.

For tagging process is simple,user tagging systems have grown in popularity on the web. However, free vocabulary has lack of standardization and semantic ambiguity. To improve the recommendation effect of labeling systems help users organize, manage and share network resources, and improve the retrieval results, labeling information based on user ontology learning methods for different mapping situations was proposed, the model of the corresponding ontology learning and semantic disambiguation model were de- signed,Ontology-based semantic tag information and ontology semantic relationships based on the extended label sorting method were designed for recommend label. Experiments show that the recall and precision have been improved,method has a good feasibility.