On a tool for rough non-deterministic information analysis and its perspective for handling numerical data

Abstract

Rough Non-deterministic Information Analysis (RNIA) is a framework for handling rough sets based concepts, which are defined in not only Deterministic Information Systems (DISs) but also Non-deterministic Information Systems (NISs), on computers. This paper at first reports an overview of a tool for RNIA. Then, we enhance the framework of RNIA for handling numerical data. Most of DISs and NISs implicitly consist of categorical data, and multivariate analysis seems to be employed for numerical data. Therefore, it is necessary to investigate rough sets based information analysis for numerical data, too. We introduce numerical patterns into numerical values, and define equivalence relations based on these patterns. Due to this introduction, it is possible to handle the precision of information, namely it is possible to define fine information and coarse information. These fine and coarse concepts cause more flexible information analysis, including rule generation, from numerical data. © Springer-Verlag Berlin Heidelberg 2005.