CAJ | 학술논문

针对当前信息检索服务中存在的固有缺陷，提出了一种基于用户桌面信息抽取的个性化推荐方法．详细介绍了通过用户桌面资源信息抽取建立长期用户模型，以及通过工作场景信息抽取建立短期用户模型的算法．长期用户模型提供了完整全面的用户兴趣偏好信息，短期用户模型则为预测用户当前信息需求提供了依据．实验结果表明，基于用户桌面信息抽取的个陛化推荐服务能较好地预测用户当前需求、具有良好的推荐效果．
침대당전신식검색복무중존재적고유결함，제출료일충기우용호탁면신식추취적개성화추천방법．상세개소료통과용호탁면자원신식추취건립장기용호모형，이급통과공작장경신식추취건립단기용호모형적산법．장기용호모형제공료완정전면적용호흥취편호신식，단기용호모형칙위예측용호당전신식수구제공료의거．실험결과표명，기우용호탁면신식추취적개폐화추천복무능교호지예측용호당전수구、구유량호적추천효과．
In terms of the inherent detects in present information retrieval service, this paper proposed an approach to personalize recommendation based on desktop information extraction and introduced the algorithm to build the long-term user model based on desktop resources extraction, which provides information about the comprehensive interest preference of a user, and to establish the short-term user model based on working scenario information extraction, which serves as the basis to predict the user＇s current information need. The experiment results showed that this approach could efficiently and effectively predict the user＇s current information need and achieved pretty good recommendation results.