Journal Article

·2008

A dynamic self-learning method for semantic query optimisation

Ayla Şaylı YTU , Oncu Uysal YTU

International Journal of Technology Policy and Management

Abstract

Semantic Query Optimisation (SQO) uses rules learned from past queries in order to execute new queries more intelligently without accessing a database, whenever possible. It has several components: Query Representation, Query Optimisation, Automatic Rule Derivation and Rule Maintenance. Automatic Rule Derivation is the main focus in this paper. A dynamic statistical learning method takes the answer set of a query, and divides it into two groups: dependent and independent attributes. Then it tests values of these attributes on whether these attributes are related or not. If they are, the method can derive new rules. Elimination is done according to the averaged rank of coefficients of linear multiple regression analysis. The method is efficient, fast and completely dynamic. It can be done on any database at any time, without any need for reconstruction on the components. Computational results of the method prove that it limits the number of rules easily and reduces the derivation time.

Keywords

Computer science Query optimization Set (abstract data type) Data mining Rank (graph theory) Representation (politics) Sargable Web query classification Web search query Machine learning Theoretical computer science Information retrieval Search engine Mathematics

Subject Areas

Advanced Database Systems and Queries ·Computer Networks and Communications ·Physical Sciences
Data Management and Algorithms ·Signal Processing ·Physical Sciences
Web Data Mining and Analysis ·Information Systems ·Physical Sciences

Citations by Year