RESEARCH ON INFORMATION INTEGRATION SYSTEM BASED ON INCOMPLETE DATABASE CONCEPT

该项目的目的是研究处理分散在Internet等开放环境中的各种数据库的信息集成系统的基础和实现技术。开放环境中的每个数据库都被认为是一个不完整的数据库，它存储了由各种数据库组成的虚拟数据库的一部分。有必要开发一种有效的方法来检索许多站点中的数据，这些站点以各种形式存储数据，包括压缩数据。搜索引擎以压缩的形式存储数据，它们使用索引来高效地搜索必要的数据。在这些系统中，同时考虑搜索效率和压缩比是很重要的。综合查询处理是基于不完备的数据库概念实现的，并且可以利用现有的数据库系统来实现不完整的数据库系统。通过一个实验系统的实现，研究了构建这样一个系统的方法。为了减少存储空间，实现对全局散乱信息的高效搜索，提出并研究了一种称为两阶段压缩的方法。该方法结合了基于索引的压缩和另一种普通的压缩方法，可以在不影响压缩比的情况下有效地检索数据。通过实现和测量两阶段压缩方法的性能来研究其性能。通过采用综合指标而不是单独的指标来提高方法的性能。利用报刊数据对该方法进行了评估，并通过可变码长对该方法进行了改进。实验还发现，当改变常用的压缩算法时，该方法的压缩比变化很大。

The aim of the project is to study the foundation and implementation techniques of information integration systems that process various databases scattered over open environment such as the Internet. Each database in an open environment is considered as an incomplete database which stores a part of a virtual database that consists of various databases. It is necessary to develop an efficient method to retrieve data in many sites which store data in various forms including compressed data. Search engines store data in compressed forms, and they use indices in order to efficiently search necessary data. It is important to consider both efficiency of search and compression ratio in these systems.Integrated query processing is realized based on incomplete database concepts, and it is possible to implement an incomplete database system using an existing database system. The method to construct such a system is investigated by implementing an experimental system. A compression method called the two stage compression method is also proposed and studied in order to reduce storage and to realize efficient search of globally scattered information. This method combines index-based compression and another ordinary compression method, and it is possible to efficiently retrieve data without compromising compression ratio.The performance of the two stage compression method is studied by implementing and measuring its performance. The performance of the method can be improved by adopting integrated indices instead of individual indices. The method was evaluated using data of newspapers and magazines, and was improved by variable code size. It is also found that the compression ratio of the method greatly varies when ordinary compression algorithm used in it is chaned.