Web数据库关键字查询结果排序方法

A ranking approach for keyword query results against web databases

为提高现有关键字查询结果的准确性,提出一种基于信息检索排序模型的关键字查询结果排序方法,该方法根据查询关键字在查询结果元组树和数据库中出现的频率来衡量查询结果的相关度,考虑了查询结果元组树中文本属性所包含字符串长度对排序分值的影响,使评分函数的计算更为合理.研究结果表明:提出的关键字查询结果排序方法的排序准确性高于现有典型排序方法,具有较高执行效率.

To deal with the accuracy of the existing keyword query results, a novel ranking approach is proposed, which is based on information retrieval ranking model. This approach takes advantages of the frequency of occurrence of keyword appeared in both joint tuple trees and database to measure the relevance degree. While the impact of length of string contained in the text attributes to the ranking score of joint tuple tree is also considered, which makes the scoring function more reasonable. Experimental results demonstrate that the accuracy of the keyword query result ranking approach is higher than that of the existing classical ranking methods, and the ranking approach proposed has higher performance as well.