An approach to flexible information access systems using soft computing

We present a scheme for modeling expert like flexibility in query-answering through extending information bases with a knowledge-based soft computing layer for query processing. The knowledge extends the each attribute with a similarity measure on its domain. The measures are applied in a soft query interpretation into a fuzzy set expression of the information need to be satisfied. The softening of query criteria applies a common model for numerical and non-numerical attributes. The query answer is a fuzzy subset of the most satisfying objects. An envelope calibration method is proposed for fast retrieval of these objects from the information base.