Jerry B. Weinberg, Ph. D.
Acting Associate Provost for Research and
Dean The Graduate School

The design of the Knowledge Discovery System focuses on the steps to be taken before the actual data mining can take place, more specifically, steps one through three as outlined in the ?KDD Overview?. Thus, the key is to implement an effective communication path between the user, the data preprocessing and transformation methods, the database, and the search algorithm. The goal of the interface is to provide a way for the user of the system to direct the KDD search process by filtering the data and sending in specific data elements from the database. In order to accomplish these tasks, it is necessary to work directly with the SQL (Standard Query Language) that the database uses, and to dynamically obtain the components of the database that is being worked on. However, the design is such that the user need not be an expert of SQL. Through the interface and selection process, the system will dynamically build the SQL statements for the user.

A Knowledge Discovery System Framework

Laurie Corwin

Abstract