基于分布式结构的数字图书馆文献集成检索方法研究  

Research on document integrated retrieval method of digital library based on distributed structure

作　　者：张小娟[1] 张永恒[1] 杨斐[1] ZHANG Xiao-juan;ZHANG Yong-heng;YANG Fei(School of Information Engineering,Yulin University,Yulin 719000,China)

机构地区：[1]榆林学院信息工程学院,陕西榆林719000

出　　处：《电子设计工程》2020年第12期35-38,43,共5页Electronic Design Engineering

基　　金：陕西省教育科学规划课题(SGH18H418);陕西省教育厅科学研究项目(18JK0909);陕西省教育科学规划课题(SGH17H282)。

摘　　要：在开放的Hadoop环境下,提出一种基于分布式结构的数字图书馆文献集成检索方法。构建五层次的分布式索引结构,合并文本文献构成集合,将分布式索引结构均匀部署于网络拓扑的节点处;采用流形排序的方式调整文献检索顺序,对输入分布式节点的指令进行赋值,计算映射函数的关系并标记待检索文献未知向量,计算权值比重,确定节点数据与被检索目标文献的相近度,完成检索方法研究。测试结果表明,提出检索方法的索引构建效率、更新效率均优于传统集中式检索方法,具有更好的文献检索效果。In the open Hadoop environment,this paper proposes a method of document integrated retrieval of Digital Library Based on distributed structure.Build a five level distributed index structure,combine text documents to form a set,and evenly deploy the distributed index structure at the nodes of the network topology;adjust the document retrieval order by manifold sorting,assign values to the instructions of the input distributed nodes,calculate the relationship of the mapping function,mark the unknown vector of the documents to be retrieved,calculate the weight proportion,and determine the node data The similarity between the retrieved target documents and the retrieval method is studied.The test results show that the index construction efficiency and update efficiency of the proposed retrieval method are better than that of the traditional centralized retrieval method,and have better document retrieval effect.

关 键 词：分布式结构 集成检索 索引结构 权值比重 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]