KNOWLEDGE REPRESENTATION AND DISCOVERY BASED ON LINGUISTIC ATOMS

An important issue in knowledge discovery in databases (KDD) is to all ow the discovered knowledge to be as close as possible to natural lang uages to satisfy user needs with tractability on one hand, and to offe r KDD systems robustness on the other. At this junction. this paper de scribes a new concept of linguistic atoms with three digital character istics: expected value Ex entropy En, and deviation D. The mathematica l description has effectively integrated the fuzziness and randomness of linguistic terms in a unified way. Based on this model, a method of knowledge representation in KDD is developed which bridges the gap be tween quantitative and qualitative knowledge. Mapping between quantiti es and qualities becomes much easier and interchangeable. In older to discover generalized knowledge from a database, we may use virtual lin guistic terms and cloud transforms for the auto-generation of concept hierarchies to attributes. Predictive data mining with the cloud model is given for implementation. This further illustrates the advantages of this linguistic model in KDD. (C) 1998 Elsevier Science B.V. All ri ghts reserved.