基于关键词提取模型与强化学习的高校图书馆图书自动分类方法  

Reinforcement Learning and Keyword Extraction-Based Automatic Book Classification in Academic Libraries

作　　者：庄辉凤 ZHUANG Huifeng(Xiamen Institute of Technology Library,Xiamen Fujian 361024,China)

机构地区：[1]厦门工学院,福建厦门361024

出　　处：《长春工程学院学报(自然科学版)》2025年第1期114-118,共5页Journal of Changchun Institute of Technology:Natural Sciences Edition

摘　　要：图书馆是高等院校的主要构成部分之一,为学生获取知识提供充足的文献资源。随着高校图书馆建设规模的扩大,文献种类与数量也呈现急剧增加的趋势,但图书分类手段依旧落后,无法满足图书的分类管理需求,故提出基于关键词提取模型与强化学习的高校图书馆图书自动分类方法。构建图书文本关键词提取模型,精准地提取图书文本关键词,以此为基础表征图书。以关键词在图书文本中出现的次数为依据,衡量关键词在图书文本中占据的权重系数,以此为依据降序排列关键词,保留固定数量的关键词,获取图书文本最终表征结果。基于强化学习理论制订图书自动分类程序,确定图书分类规则,接下来执行制订程序即自动可获取最终的图书分类结果。试验数据显示:在不同试验组别背景下,提出的方法被应用后获得的图书分类敏感度最大值为96%,图书分类几何平均值最大值为92%,充分证实了提出方法的图书分类效果更佳。The library is one of the main components of higher education institutions,providing sufficient book resources for students to acquire knowledge.With the expansion of the construction scale of university libraries,the types and quantities of books also show a sharp increase trend,but the means of book classification is still backward,unable to meet the needs of book classification management,so this paper proposes a research on automatic classification methods of university library books based on keyword extraction model and reinforcement learning.Build a book text keyword extraction model,accurately extract book text keywords,and represent books based on this.Based on the number of occurrences of keywords in the book text,the weight coefficient of keywords in the book text is measured.Based on this,keywords are arranged in descending order,retaining a fixed number of keywords to obtain the final representation result of the book text.Based on the theory of reinforcement learning,the automatic book classification program is developed,and the rules of book classification are determined.The final results of automatic book classification can be obtained by executing the program.The experimental data shows that under different experimental group backgrounds,the maximum sensitivity of book classification obtained after the application of the proposed method is 96%,and the maximum geometric average of book classification is 92%,fully confirming that the proposed method has better book classification performance.

关 键 词：强化学习 高校图书馆 关键词提取模型 图书分类 相似度计算 图书管理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]