A KNOWLEDGE-BASED APPROACH TO STATISTICAL QUERY-PROCESSING

Representation of statistical data is achieved by means of complex data structures. Such a complexity transfers directly to statistical data manipulations. Furthermore, the most meaningful statistical indicator resides in complex data structures. Our work is focused on simplifying manipulations of statistical data structures by means of StEM (Statistical database Expert Manager), a knowledge-based kernel. StEM acts as a deductive query processor between a high level query language and the set of low level DBMS primitives required for query resolution. StEM contains a Prolog knowledge base that provides transparency of complex data manipulations, allowing a real conceptual interaction with a statistical database.