Allowing Users to Create Similarity Relations for Their Flexible Searches over Databases

Abstract

A bi-valued logic is not enough to make an intelligent search engine to give us the result for the queries like “I am looking for a cheap restaurant, Mediterranean food or similar type”. With the integration of Fuzzy Logic and Logic Programming, we were able to model and pose flexible queries over databases. Therefore, we present a framework that allows users to pose their expressive queries based on defining similar relation criteria over various modern and conventional data formats such as JSON, SQL, CSV, XLS, and XLSX. The interest is in, for example, obtaining “drama movie” when asking for “romantic movie” (only if the similarity relation between drama and romantic movie is explicitly defined in the configuration file). The uses of similarity relation between values allow us to obtain more answers apart from the identical one. The searches that use two or more criteria are much more expressive and accurate. This framework provides the facility to define, modify and remove similarity relations from a user-friendly interface (without the need to be concern about the low-level syntax of the similarity criteria).