Abstract: At present most personalization service systems use keywords or keyword vector space models to express users’ personal preferences, but it severely restricts the sharing and reusing of user model between different personalized systems. Focusing on the problem, this paper proposes a user model called DOBUM(Domain Ontology-Based User Model) and presents its structure. To realize the quantification of users’ personal preference, the definition of personalized interest level is introduced. The application process of DOBUM in the group user modeling is presented. The model is proved to reflect users’ interests exactly and have strong renewal ability and extendibility by analyzing the contrastive result on data.

Key words: cross system personalization service, ontology, domain ontology, user model

摘要： 目前大多数个性化服务系统采用关键词或关键词向量空间模型表示用户个人偏好，该方法严重制约了用户模型在不同系统间的共享和重用。针对上述问题，提出一种基于领域本体的用户模型，给出其形式化描述，通过引入个性化兴趣度实现用户个人偏好的量化，介绍该模型在群体用户建模中的应用过程。实验表明，该用户模型能准确反映用户兴趣且更新能力和可扩展性较强。

关键词: 跨系统个性化服务, 本体, 领域本体, 用户模型

CLC Number: 

• N945