The most important objective of this study is to develop a mete-level knowledge base system for realizing a creative environment of new research fields by integrating information resources in cultural, social, and natural sciences. The meta-level knowledge base system realizes highly creative activities for human beings over various research fields by sharing, retrieving, editing and integrating databases through wide area computer networks.

A number of legacy databases for individual scientific research fields are connected to wide area computer networks. As the existing legacy databases for those research fields have been designed and created with their own data structures, data representations and languages, it is difficult to obtain global knowledge by sharing, retrieving, editing and integrating those databases. In such a heterogeneous database environment, semantic heterogeneity poses problems in integrating different databases. Knowledge sharing, semantic retrieving, editing and integrating for those databases are essentially important for creating new research fields dynamically over various research fields, and the meta-level knowledge base system can realize an intelligent knowledge integration environment for future scientific research.

[A System Architecture of Meta-Level Knowledge Base Systems]

In this study, we design and develop a meta-level knowledge base system and realize an intelligent knowledge base environment for creating new research field with existing various information resources. In this system, databases for cultural, social and natural sciences are connected to the meta-level layer, and those databases are integrated by temporal, spatial and semantic functions. By the connection and integration among those databases from different fields, this system provides a knowledge integration environment for new scientific research related to various scientific fields.

The meta-level system is located in the meta-level layer of the existing local and legacy databases which include existing information resources from various scientific fields. According to queries, this system performs retrieving, editing, integrating and mining databases by translating data structures, data representations and language representation into the common data structures and representations, and by applying temporal, spatial and semantic functions to those data.

We have presented a meta-database system architecture and its implementation model. We have also developed several query processing environments in the mete-database system based on the proposed model. Furthermore, we have designed and developed a semantic associative search method which computes dynamic and semantic correlations with a context recognition mechanism. We have applied this method to several multimedia database applications, such as image and music data retrieval by human impression. This research results have widely been introduced in the book of ``Multimedia Data Management - using metadata to integrate and apply digital media --, McGraw Hill (book), Chapter 7, 1998.'' We create a new knowledge base system environment by applying those methods to data retrieval, data integration and data mining.


[top]