Contextual Ranking of Database Querying Results: A Statistical Approach

摘要

There has been an increasing interest in context-awareness and preferences for database querying. Ranking of database query results under different contexts is an effective approach to provide the most relevant information to the right users. By applying the regression models developed in the statistics field, we present a quantitative way to measure the impact of context upon database query results by means of contextual ranking functions with context attributes and their influential database attributes as parameters. To make the approach computationally efficient, we furthermore propose to reduce the dimensionality of context space, which can not only increase computational efficiency but also help ones identify informative association patterns among context attributes and database attributes. Our experimental study on both synthetic and real data verifies the efficiency and effectiveness of our methods.