摘要
提出了一种基于搜索代价的对Web数据库多查询结果进行分类的方法,该方法首先通过分析用户的查询习惯,构建一个通用的查询结果分类树探测模型,然后根据探测模型建立分类树的搜索代价模型。对于搜索代价,提出了基于查询历史的搜索代价估计方法。最后,以降低搜索代价为目标在查询结果集上生成一个分类树,用户通过检查该分类树上各分支节点的标签来逐步定位到其感兴趣的信息。实验及分析表明,本文所提方法能够有效避免信息过载,并且具有较好分类效果和较低搜索代价。
This paper proposes a categorization approach to query results based on searching cost. Firstly, a general exploration model which meets users' query habits is presented. And then, a searching cost model is built corresponding to the exploration model. To estimate the searching cost, this paper proposes a searching cost measuring method by taking advantage of query history. Lastly, a labeled and leveled categorization tree is generated according to the searching cost. By using the categorization tree, users can easily find their favorite results by checking the label assigned on the tree nodes. The experiments demonstrate that the method can efficiently avoid the information overload, and has the higher categorization accuracy and lower searching cost as well.
作者
高建
GAO Jian(Department of Mechanical and Electrical Engineering, Panjin Vocational & Technical college, Panjin 124010, China)
出处
《辽宁工业大学学报(自然科学版)》
2017年第2期85-90,共6页
Journal of Liaoning University of Technology(Natural Science Edition)
关键词
搜索代价
信息过载
查询结果分类
searching cost
information overload
query result categorization