CPrefSQL Extension

Introduction

The widespread of applications like e-commerce and recommendation systems requires the use of efficient customization techniques and user preferences handling. In database community a lot of research on this topic has been focused on extending standard SQL with preference facilities in order to provide personalized query answering. More specifically, user preferences are an essential ingredient of personalized database applications.

The CPrefSQL language was created to provide support to queries with conditional preferences. The conditional preferences allow users to express how their wishes over an attribute are affected by the value of other attributes. When a CPrefSQL query (or cp-query) is processed the tuples of a relation are compared according to the conditional preference rules (or cp-rules) specified by the user. Actually, the cp-queries can be used to get the dominant tuples or the top-k tuples. A tuple t is dominant if there no exists another tuple better than t in the relation. The top-k tuples are the tuples having the lowest preference level. This project concerns to develop new kinds of queries to CPrefSQL.

Current Implementation

Available at: http://cprefsql.github.io/

Publications

• RIBEIRO, M. R.; PEREIRA, F. F. S. ; DIAS, V. V. S. (2016). Efficient Algorithms for Processing Preference Queries. In: Annual ACM Symposium on Applied Computing (SAC), Pisa, Italy.
• de AMO, S. ; RIBEIRO, M. R. (2009). CPref-SQL: A Query Language Supporting Conditional Preferences. In: Annual ACM Symposium on Applied Computing (SAC), Honolulu, USA.