关键词: 数据源选择, 关键词查询, 概要, min-hash 算法

Abstract: With the rapid growth and deployment of the distributed databases over the Internet, it calls for new efficient search method over multiple structured data sources. This paper proposes a new keyword-search method for effective database selection using inverted lists. The method can achieve a high interactive speed and thus can improve user experiences. This method has been implemented on airticket-search systems, and experimental results show that it achieves high search performance. For large scale data, a min-hash based algorithm is adopted to select highly relevant data sources, which can improve the performance and achieve high precision

Key words: database selection, keyword search, database summary, min-hash based algorithm