基于特征索引的图相似查询过滤算法

Graph Similarity Query Filtering Algorithm Based on Feature Index

分析图相似查询候选集的产生过程以及特征图之间的关系对候选图集的影响,提出一种基于特征索引的图相似查询过滤算法,使用GIndex算法建立特征图索引结构,通过特征图之间的选择性关系给出一个有序的特征集,并借助特征-图矩阵对数据库进行筛选得到候选图集。实验结果证明,该方法能准确地产生候选图集,从而提高图查询的效率。

This paper analyzes the generation process of the graph similarity search candidates, and the relationship between feature-graph that not be token into account in current filtering algorithm, and presents graph similarity query filtering algorithm based on feature index. It builds a feature-graph index structure, analyzes the selectivity of each feature and builds a sorted feature, and generates the candidate graphs by using the feature-graph matrix to efficiently filter the graph database. Experimental results show that this algorithm Can accurately generate candidate and improve the efficiency of the graph query.