数据仓库系统技术研究

This project focus on the implementation techniques of parallel data warehouse, including framework, storage organization, cache mechanism, three level index technique, transaction processing and query optimization, and so on. Especially, we have went deep into the cube computation, and proposed a cube compressing technique, a parallel computing algorithm, and a cube selection algorithm. We have also studied the possibility of extending data warehouse management function on parallel database system PBASE, and developed a prototype named PARAWARE. In addition, we have studied the issues of data mining and information management on Internet, and proposed a method that user constraint is used to decrease the dimensional space when probe data cube in warehouse. We also proposed algorithms for text classification and feature selection.

本项目的研究目标是对数据仓库管理技术（数据抽取、组织策略、存储优化和维护燃际酰┮约盎谑莶挚獾牟檠呕呗裕ň奂檠⒋笫萘坎檠⒎植疾檠鹊挠呕┑难芯浚愿行У刂С諳LAP和数据挖掘应用；同时结合我研究所已研制的－－并行数据庀低砅ABSE（国家‘863’科研成果），为在国产数据库系统软件上扩展数据仓库管理功能投圆咧С钟τ玫闹С肿隼砺垩芯浚⑻峁┦笛橄低场