针对目前内容过滤算法的不足和数字家庭业务特点,提出了基于内容过滤的数字家庭服务资源推荐技术。采用三元组构建用户偏好模型,将用户偏好特征从资源的文字介绍信息具体到每一个属性,并引入遗忘因子动态挖掘用户偏好,自适应调节特征权重,并根据服务资源模型与用户偏好模型的相似性向用户返回推荐列表。实验结果表明,基于内容过滤的推荐方法可使用户点播率和反馈率得到提高。
For the personalized demands of digital home users, the content - based filtering recommendation technique was proposed for users to get valuable service resources. The ternary group was used to build user preference model. User preference characteristics were specified from service resources introductory information to their each attribute; and forgetting factor was introduced to dynamically mine user preferences. The weight of characteristics was self - adaptively adjusted. Finally according to the similarity between service resources and user preference model, recommendation list was returned to users. The experiment results show that content -based filtering recommendation method makes users'demand rate and response rate increase.