基于Spark框架的图书馆文献信息检索方法  

Library Literature Information Retrieval Method Based on the Spark Framework

作　　者：翟小乐 任云鹏[1] 蒋丽铭 ZHAI Xiaole;REN Yunpeng;JIANG Liming(Qingdao Vocational and Technical College of Hotel Management,Qingdao Shandong 266000,China)

机构地区：[1]青岛酒店管理职业技术学院,山东青岛266000

出　　处：《信息与电脑》2023年第8期60-62,共3页Information & Computer

摘　　要：由于传统信息检索方法检索时间长,检索误差率高,研究基于Spark框架的图书馆文献信息检索方法。首先,预处理文献信息数据,结合信息相似度,提取特征数据集合中的敏感数据获得属性特征。其次,基于Spark框架批处理数据,训练待测样本进行目标跟踪,寻找最优解集。对信息的特征进行查询,提取其中的关键词并在已建立的文本索引结构上进行检索和融合,运用K-Means聚类方法对融合结果进行重新排序。采用布尔查询机制构造文献检索模型,将不同信息数据多重倒排文献进行索引从而完成信息检索。实验结果表明,运用该方法检索时间最短,误差率最低,完成图书馆文献信息检索方法的优化。Due to the long retrieval time and high retrieval error rate of traditional information retrieval methods,this paper studies the library document information retrieval method based on Spark framework.Firstly,the literature information data is preprocessed,and the sensitive data in the feature data set is extracted according to the information similarity to obtain attribute features.Secondly,based on Spark framework batch processing data,the samples to be tested are trained for target tracking to find the optimal solution set.The features of the information are queried,the keywords are extracted and retrieved and fused on the established text index structure.The fusion results are reordered by K-Means clustering method.A literature retrieval model is constructed by using Boolean query mechanism to index multiple inverted literature with different information data to complete information retrieval.The experimental results show that the method has the shortest retrieval time and the lowest error rate,and completes the optimization of library literature information retrieval method.

关 键 词：Spark框架 图书馆 信息检索 文献 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]