Knowledge representation and discovery based on linguistic atoms

摘要

An impcn-tant issue in knowlcd2c discovery in databases (KDD) is to allow the discovered knowledge to be as close as possible to natural languages to satisly user needs with tractahility on one hand, and to offer KDD Systems rohustncss oil the other. At this junction. this palm descrihes a new concept of linguistic atoms with three di(,ital characteristics: expected value Ex, entropy En and deviation D. The mathematical 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 2ap hetween quantitative and yualitatM. knowledge. Mapping hetween quantities and qualities hecomes much easier and interchangeable. In order to discover generalized knowledge from a datahase, we nrty use virtual linguistic terms and cloud translortn.s for the auto-oeneration of concept hierarchies to attrihutes. Predictive data mining with the Cloud model is 2iven for implementation. This further illustrates the advantages of this linguistic model in KDD.