In this thesis we describe the UQoRE method which supports database reverse engineering by using a data mining technique. Generally, Reverse Engineering methods work by using information extracted from data dictionaries, database extensions, application programs and expert users. The main differences between all these methods rely on the assumptions made on the a-priori knowledge available about the database (schema and constraints on attributes) as well as the user competence. Most of them are based on the attribute name consistency. This paper presents a method based on user queries. Queries are stored in a “Query Base” and our system mines this new source of knowledge in order to discover hidden links and similarity between database elements.
