摘要
超级节点网络中的超级节点可能成为网络性能的瓶颈并影响检索结果的统一排序,针对该问题提出一种并行查询和排序机制。给出类特征等索引建立方法和查询节点选择算法,减少超级节点的存储和计算负担,使其在负载能力范围内,尽可能多地连接普通节点。在获取全局参数的前提下,提出查询节点的查询和排序方法,以提高检索质量。实验结果验证了该机制的有效性。
Super peers in super-peers network may become the network performance bottleneck and affect retrieval result global ranking. Aiming at this problem,this paper presents a parallel query and ranking mechanism. A building method for indexes such as class feature and a selection algorithm for query peers are presented to decrease the storage and computation burden of super peers and make it connect more common peers as possible in the scope of load capacity. A query and ranking method of query peers is presented to improve retrieval quality. Experimental results show that the mechanism is efficient.
出处
《计算机工程》
CAS
CSCD
北大核心
2010年第2期97-99,共3页
Computer Engineering
基金
湖南省教育厅科研基金资助项目(07B007)
关键词
超级节点网络
信息检索
并行查询
super-peers network
information retrieval
parallel query