Data clustering technique has been applied in many research areas. An efficient data clustering mechanism not only discriminates patterns from an information system but also increases the visibilities of the recognized patterns. In general, a decision table in an information system is the kernel of the decision-making processes that generate decision rules. Considering an information system, the object and the attribute are two essences that construct a decision table. The attribute values also called features, which describe object behaviors, have various domain types. In practice, a typical attribute domain could contain either a set of continuous numbers or a set of nominal codes. The variety in attribute domains increases computational complexity and difficulty. In some research areas, such as data mining and knowledge discovery from database(KDD), a meaningful and easy-to-interpret information model is required. Thus, an effective data transformation method or an efficient data clustering mechanism for those information models are highly in demand. In this paper, we proposed an effective data transformation approach that precisely partitions continuous numbers into optimized discrete intervals. The proposed technique has been applied to different sets of data for comparison. The result of our experiments is encouraging.
DOI 인용 스타일