论文部分内容阅读
为了解决用户兴趣建模初期存在的冷启动问题,以科研用户公开发表的学术产出作为用户兴趣建模的数据源,采用文本挖掘和基于本体的模型表示技术进行用户建模,并提出一种通过实体关系表示用户兴趣的方法。该方法与使用单个关键词或实体的表示方法相比,语义信息更为丰富,能更好地描述用户兴趣。最后,将生成的用户兴趣本体实例存储到Sesame本体数据库中,支持通过SeRQL和SPARQL语言进行查询,实现了用户兴趣信息的语义化存储和检索。
In order to solve the problem of cold start at the beginning of user interest modeling, the academic output published by scientific research users is used as the data source of user interest modeling. The text mining and ontology-based model representation techniques are used to model users. A method of expressing user interest through an entity relationship. Compared with the representation method using a single keyword or entity, the method is more semantic information and can better describe the user’s interest. Finally, the generated instances of user interest ontology are stored in the Sesame Ontology database, which supports query through SeRQL and SPARQL to achieve semantic storage and retrieval of user interest information.