On Modeling Query Refinement by Capturing User Intent through Feedback
SQL queries in relational data model implement the binary satisfaction of tuples. Tuples are generally filtered out from the result set if they miss the constraints expressed in the predicates of the given query. For naïve or inexperienced users posing precise queries in the first place is very difficult as they lack of knowledge of the underlying dataset. Therefore, imprecise queries are commonplace for them. In connection with it, users are interested to have explanation of the missing answers. Even for unexpected tuples present in the result set advanced users may also want to know why a particular piece of information is present in the result set. This paper presents a simple model for generating explanations for both unexpected and missing answers. Further, we show how these explanations can be used to capture the user intent via feedback specifically for refining initial imprecise queries. The presented framework can also be thought as a natural extension for the existing SQL queries where support of explanation of expected and unexpected results are required to enhance the usability of relational database management systems. Finally, we summarize future research directions and challenges that need to be addressed in this endeavour.
Proceedings of the Twenty-Third Australasian Database Conference (ADC 2012)
Pattern Recognition and Data Mining