Database Preferences -A Unified Model
DocUID: 2012-005 Full Text: PDFAuthor: Roxana Gheorghiu, Alexandros Labrinidis, Panos K. Chrysanthis
Abstract: In this work we present a new model that combines two different types of preferences, qualitative and quantitative. We show how our model can support different types of preferences at different granularity levels and how an application can use these preferences to retrieve a list of tuples. The new model takes advantage of a graph representation of preferences where nodes in the graph are SQL predicates, edges between two nodes describe a qualitative preference and edges from a node to the same node capture quantitative preferences. Each edge is labeled with a numeric value, between -1 and 1, that expresses the intensity of each preference. Using this graph representation, we further show how two different preferences can be combined and applied to the existing query result set to filter the result and identify the most relevant tuples first.
Keywords: preferences, database
Published In: 6th International Workshop on Personalized Access, Profile Management, and Context Awareness in Databases
Pages: 1-6
Year Published: 2012
DOI: http://persdb2012.cs.umn.edu/papers/3.Gheorghiu-PersDB12.pdf
Project: Others Subject Area: Database Preferences
Publication Type: Conference Paper
Sponsor: NSF OIA-1028162, NSF IIS-0952720, NSF IIS-0746696