Studies on Stream Mining in Web Archive

网络档案的规模呈指数级增长，许多国家图书馆和IIPC（国际互联网保存联盟）正在努力制定长期保存数字内容的指导方针。本文从声誉模型数据挖掘技术的角度出发，重新考虑了web归档系统存储量的增长模型。基于RAM、HDD、磁带和磁盘等存储设备的特点，讨论了分层存储系统的基本架构。利用文件检索模式和访问频率对文件移动算法进行了改进。

The size of the web archive is increasing exponentially, many national libraries and IIPC (International Internet Preservation Consortium) are making efforts to decide guidelines of long-term preservation of digital contents. In this research, from the view points of data mining techniques for reputation model, we reconsider a growth model of storage volume in web archive system. We discuss a basic architecture of hierarchical storage system based on characteristics of memory devices such as RAM, HDD, magnetic tapes and disks. We improve the file moving algorithm by using file retrieval patterns and access frequencies.