personalized search and recommendations for movies in a relational database
S.Arun Raj,V. Loganathan
Published in International Journal of Advanced Research in Computer Science Engineering and Information Technology
ISSN: 2321-3337 Impact Factor:1.521 Volume:4 Issue:3 Year: 06 May,2016 Pages:1031-1038
Abstract
PrefDB, a preference aware relational system that transparently and efficiently handles queries with preferences. In its core, PrefDB employs a preference aware data model and algebra, where preferences are treated as first class citizens. We define a reference using a condition on the tuples affected, a scoring function that scores these tuples, and a confidence that shows how confident these scores are. In our data model, tuples carry scores with confidences. Our algebra comprises the standard relational operators extended to handle scores and confidences. For example, the join operator will join two tuples and compute a new score confidence pair by combining the scores and confidences that come with the two tuples. In addition, our algebra contains a new operator, prefer, that evaluates a preference on a relation, i.e., given as inputs a relation and a preference on this relation, prefer outputs the relation with new scores and confidences. During preference evaluation, both the conditional and the scoring part of a preference are used. The conditional part acts as soft constraint that determines which tuples are scored without disqualifying any tuples from the query result. In this way, PrefDB separates preference evaluation from tuple filtering. This separation is a distinguishing feature of our work with respect to previous works. It allows us to define the algebraic properties of the prefer operator and build generic query optimization and processing strategies that are applicable regardless of the type of reference specified in a query or the expected type of answer
Kewords
Database personalization, preferences
Reference
. G. Adomavicius and A. Tuzhilin (Jun. 2005), Toward the next generation of recommender systems A survey of the state-of-the-art and possible extensions, IEEE Trans. Knowl. Data Eng., vol. 17, no. 6, pp. 734 749. 2. R. Agrawal, R. Rantzau, and E. Terzi, (2006) Context-sensitive ranking, in Proc. SIGMOD, Chicago, IL, USA, pp. 383 394. 3. R. Agrawal and E. L. Wimmers, (2000) A framework for expressing and combining preferences, in Proc. SIGMOD, Dallas, TX, USA, pp. 297 306.
