SIMPLE DATA FILTERING IN ROUGH SET SYSTEMS

In symbolic data analysis, high granularity of information may lead to rules based on a few cases only for which there is no evidence that t hey are not due to random choice, and thus have a doubtful validity. W e suggest a simple way to improve the statistical strength of rules ob tained by rough set data analysis by identifying attribute values and investigating the resulting information system. This enables the resea rcher to reduce the granularity within attributes without assuming ext ernal structural information such as probability distributions or fuzz y membership functions. (C) 1998 Elsevier Science Inc.