基于图的音乐数据查询处理及优化方法

An Approach to Music Data Query Processing and Optimization Based on Graph

音乐数据之间存在复杂关联关系,适合用图数据结构对其建模并进行查询处理.然而现有的图查询处理算法只关注图结构特征,并未针对音乐元数据和音乐内容数据进行优化,查询效率不高.基于图音乐数据模型GraMM与查询语言GraMQL,提出了基于图的音乐数据查询处理算法.该算法根据音乐数据的特点,使用图结构剪枝、音乐元数据剪枝以及音乐内容剪枝3种策略对搜索空间进行剪枝,提高了查询效率.进而给出了调整查询顶点搜索顺序的优化方法以及基于开销模型的音乐内容剪枝位置优化方法,加快了查询处理速度.实验结果表明所提音乐查询处理及优化算法能高效处理音乐元数据和音乐内容数据查询请求.

Considering the complicated relationship among music data,it is very natural to model music data as graph.However existing graph query algorithms cannot efficiently process queries since there is no optimization for music metadata and music content.Based on the proposed graph music data model GraMM and query language GraMQL,we first present a query processing algorithm which exploits graph structure,metadata and content data of music to prune search branches.There are three pruning strategies employed in our approach.Then a query optimization method is given to rank search order and find proper position for content pruning based on a cost model.Finally the experimental results show that our methods are effective and efficient for querying on both music metadata and music content.