论文部分内容阅读
互联网用户偏好本体可以全面、准确地描述出互联网用户的兴趣和多维偏好。针对偏好本体中主题类的实例对象数量众多、不断扩展变化、手工搜集工作量大这一问题,重点研究用户偏好本体中主题专业网站、品牌和体育赛事三类具有代表性的实例学习方法,以期实现互联网用户偏好本体的半自动构建,并设计实验验证这三类实例学习方法的有效性。
Internet user preference ontology can fully and accurately describe Internet users’ interests and multi-dimensional preferences. In view of the large number of instance objects of the subject class in the preference ontology and the ever-expanding changes and the heavy workload of manual collection, this paper focuses on the three typical case-based learning methods of subject professional websites, brands and sports events in user preference ontology The semi-automatic construction of Internet user preference ontology is realized, and the effectiveness of these three types of instance learning methods is verified by experiments.