基于语义计算的查询扩展优化研究

Optimization Research for Query Expansion Based on Semantic Computation

查询扩展技术是指在原有查询的基础上加入与用户输入的检索用词相关联的新词,组成新的更长、更准确的查询,用于弥补用户查询信息不足的缺陷.为了提高文本检索的效率,纳入网络检索环境下的用户个人偏好,在查询扩展技术中引入语义计算是一个重要研究方向.文章从语义计算的角度提出了基于语义关联树的查询扩展算法,通过动态生成语义关联树,有效降低词相似度矩阵计算工作量.通过控制语义关联树的层次结构及复杂度,灵活高效的生成不同语义空间模型.实验证明,该算法能有效提高文本检索的准确率.

QE （Query Expansion） is to add some new correlated words to original retrieval, and make a new efficient query The main purpose of QE is to offset the defect of the user＇ s short query information. To improve text retrieval efficiently with user＇ s preference, introducing semantic computation to query expansion is a main research direction. This paper provides a new query expansion algorithm .based on semantic relation tree in the view of semantic computation. By implementing the tree dynamically, we can effectively reduce the complexity of getting words similarity matrix. By controlling the tree structure and complexity, we can effectively and neatly construct different semantic space models. The experimental result shows that this algorithm can improve the accuracy of text retrieval efficiently.