Automated knowledge acquisition of case-based semantic networks for interactive enhancement of the data mining process.

摘要

This work presents a knowledge acquisition algorithm for the automated construction of a case-based semantic network model from conventional relational databases. An associative retrieval algorithm is also provided to support interactive use of the model. Scalability of the performance of the knowledge acquisition algorithm is investigated with a parallel version. In addition to acquiring explicit information, which is represented as cases, the knowledge acquisition algorithm also captures the semantic and associative relationships which are implicit in the database. The semantic links define set membership for the acquired cases and the associative relationships define concepts. A user friendly graphical user interface has been developed to facilitate access to the knowledge base. A browsing facility, which supports navigation of semantic links, is particularly suited for enhancing interactivity of the data mining process with original data after interesting patterns have been identified.