首页 | 本学科首页   官方微博 | 高级检索  
     

个人数据空间管理中的任务挖掘策略
引用本文:寇玉波,李玉坤,孟小峰,张相於,赵婧. 个人数据空间管理中的任务挖掘策略[J]. 计算机研究与发展, 2009, 46(Z2)
作者姓名:寇玉波  李玉坤  孟小峰  张相於  赵婧
作者单位:中国人民大学信息学院,北京,100872
基金项目:国家自然科学基金项目,国家"八六三"高技术研究发展计划基金项目,教育部博士学科点专项基金项目 
摘    要:在个人数据空间管理过程中,用户需要处理大量异质数据如邮件、文档、图片等.随着用户数据在数量和种类上的增多,如何有效管理这些数据,为用户提供有效的存储及查询服务成为一个具有挑战性的问题.传统数据管理工具如文件系统、桌面搜索工具等并未给用户提供足够的管理能力.究其原因,个人数据空间是由数据、用户以及服务三要素组成.而传统数据管理工具却忽略了用户这一要素,因此仅能在存储路径或全文索引的基础上提供服务.实际上数据与用户之间具有密不可分的联系,个人数据空间中的数据正是来自于用户行为.而用户行为是由一个一个任务组成的.挖掘个人数据中的任务,可以建立起数据间基于用户行为的语义关系,进而可以为用户提供任务视角的数据管理服务以及基于任务的查询服务.正是基于这一思想,提出了基于用户行为挖掘用户任务的方法.通过分析用户行为,发现个人数据由用户行为产生的时序关系,然后根据该时序关系生成用户的任务.实验证明该方法是有效的.

关 键 词:任务挖掘  用户行为

A Task Mining Strategy in Personal Dataspace Management
Kou Yubo,Li Yukun,Meng Xiaofeng,Zhang Xiangyu,Zhao Jing. A Task Mining Strategy in Personal Dataspace Management[J]. Journal of Computer Research and Development, 2009, 46(Z2)
Authors:Kou Yubo  Li Yukun  Meng Xiaofeng  Zhang Xiangyu  Zhao Jing
Abstract:In personal dataspace management,users need to manage heterogeneous data such as emails,documents,pictures and so on.As user data rapidly increase in amount and variety,it has become a challenging problem to effectively manage these data and provide users with effective store and query service.Traditional data management tools such as file system and desktop search tools are not able to satisfy users.Personal dataspace is composed of three factors(data,user and service),while those old ones neglect the user factor.Thus they can only provide service based on paths or fulltext index.Actually,there is close relationship between data and user.Personal data comes from user behavior.Meanwhile,user behavior is composed of many tasks.Mining tasks in user data could help establish semantic associations between data and provide users with task-based data management and query service.Therefore,a method is proposed to mine tasks based on user behavior.firstly user behavior is analyzed,and time sequence relationships between personal data are found,then user tasks are mined.The experiments show this method is effective.
Keywords:task mining  user behavior
本文献已被 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号