2. 제2주제: Data Mining in Very Large Data Sets

- 발 표 자: 김진화 교수 (서강대학교, 경영대학 MIS 전공)

- 발표 내용:
Many data mining algorithms are not capable of working effectively with the very 
large datasets that organizations today are building with the massive amounts of 
Internet-related data they collect and store. This study constructs a new 
classification algorithm and methodology, referred to in this manuscript as the 
Iterative Refinement Algorithm (IRA), that builds domain knowledge from very large 
databases with an inductive learning system. Unlike existing systems that build 
complete domain knowledge from a dataset, IRA builds the domain knowledge from the 
dataset and then iteratively improves, sharpens, and polishes it, a process that 
we refer to as refinement of the domain knowledge base.