一种基于桌面信息的个性化推荐方法  

An Approach to Personalize Recommendation Based on Desktop Information Extraction

作　　者：李晓昀[1] 余颖[1] 阳小华[1] 万亚平[1] 马家宇[1] 刘志明[1] 蒋辉[1] 

机构地区：[1]南华大学计算机科学与技术学院,湖南衡阳421001

出　　处：《南华大学学报（自然科学版）》2012年第1期54-57,67,共5页Journal of University of South China：Science and Technology

基　　金：湖南省自然科学基金资助项目(10JJ6097);湖南省教育厅科研基金资助项目(11C1098);湖南省科技计划基金资助项目(2010GK3011);南华大学校级教改课题基金资助项目(2010ZZ030)

摘　　要：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性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.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.

关 键 词：桌面 个性化推荐 信息抽取 个人信息空间 

分 类 号：TP393[自动化与计算机技术—计算机应用技术] TP311[自动化与计算机技术—计算机科学与技术]