图书馆数字文本智能聚类个性化推荐应用研究  

Research on the application of intelligent clustering and personalized recommendation of digital text in libraries

作　　者：江新姿[1] 高尚 JIANG Xinzi;GAO Shang(Library of Jiangsu University of Science and Technology,Zhenjiang 212004,China;School of Computer Science,Jiangsu University of Science and Technology,Zhenjiang 212004,China)

机构地区：[1]江苏科技大学图书馆,江苏镇江212004 [2]江苏科技大学计算机学院,江苏镇江212004

出　　处：《无线互联科技》2025年第2期107-111,120,共6页Wireless Internet Science and Technology

基　　金：国家自然科学项目,项目名称:融合样本先验分布信息的类别不平衡学习理论与算法研究,项目编号:62176107;镇江社科联项目,项目名称:数字化技术赋能提升镇江市公共文化服务水平研究,项目编号:2024YBL142。

摘　　要：Web 2.0信息时代,信息量迅速增加,信息检索速率却显著降低,如何提高信息的自动分类管理水平,从海量数据中高效、准确、快速获取有价值的信息与知识成为智慧图书馆亟待研究与解决的问题。文章提出了在数字图书馆服务中运用新型文本聚类群智能分析方法。该算法通过改进文本间的语义相似度计算,融合K-means聚类算法与蚁群聚类算法(Ant Colony Optimization,ACO)的优点,在初始分类时将K-means聚类算法用作快速分类,用分类结果指导更新蚂蚁各途径信息素,指导蚂蚁后续聚类途径选择,提高聚类运行效率。该分析方法因为不需要类别的信息,能自动完成文本分组,所以可以更好地应用到图书馆资源的推荐与检索服务中。图书馆数字文本数据库实验证明,混合蚁群聚类算法比单独的K-means、ACO都具有更好的聚类效果,可以看出该算法的有效性。In the era of Web 2.0 information,the amount of information is rapidly increasing,but the speed of information retrieval is significantly decreasing.How to improve the automatic classification management of information and efficiently,accurately,and quickly obtain valuable information and knowledge from massive data has become an urgent problem for smart libraries to research and solve.The article proposes the application of a new text clustering group intelligent analysis method in digital library services.This algorithm improves the calculation of semantic similarity between texts,integrates the advantages of K-means clustering algorithm and ant colony clustering algorithm,and uses K-means algorithm for fast classification in the initial classification.The classification results guide the updating of pheromones in various ant pathways,guide the selection of subsequent clustering paths for ants,and improve the efficiency of clustering operation.Due to the fact that this analysis method does not require category information and can automatically complete text grouping,it can be better applied to the recommendation and retrieval services of library resources.Through experiments on digital text databases in libraries,it has been proven that the hybrid ant colony clustering algorithm has better clustering performance than individual K-means and ACO,indicating the effectiveness of the algorithm.

关 键 词：文本聚类 K-MEANS聚类 混合蚁群聚类算法 个性化推荐 语义相似度 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程] G251.4[自动化与计算机技术—控制科学与工程]