On fuzzy queries with contextual predicates

This paper deals with the interpretation and processing of database queries with preference conditions of the form ldquoattribute is low (resp. medium, high)rdquo in the situation where the user is not aware of the actual content of the database but still wants to retrieve the best possible answers (relatively to that content). An approach to the definition of the terms ldquolowrdquo, ldquomediumrdquo and ldquohighrdquo in a contextual and relative manner is introduced. A processing algorithm aimed at efficiently retrieving the top-k answers to such a query is also outlined.