基于机器学习的中文图书智能分类实证研究  

An Empirical Study on Intelligent Classification of Chinese Books Based on Machine Learning

作　　者：夏丹[1] Xia Dan

机构地区：[1]哈尔滨理工大学图书馆,黑龙江哈尔滨150081

出　　处：《新世纪图书馆》2025年第2期45-52,共8页New Century Library

摘　　要：为提高图书审校效率,论文以高校图书馆馆藏中文书目为数据源,以内容提要、主题词和题名为特征词来源,根据特征词来源位置对特征词进行加权处理和特征词词频统计,构建图书-特征词稀疏矩阵,按比例对带有图书分类号的稀疏矩阵进行朴素贝叶斯计算,找到图书分类最大概率,评估训练分类模型。实验结果表明,利用朴素贝叶斯算法基于加权精选特征词的图书智能分类模型,具有良好的实用性,对进一步提高采编部工作的智能化和高效化是有效可行的。To improve the efficiency of book review,this paper takes the Chinese bibliography collected by university library as the data source,takes the content summary,subject words and titles as the source of feature words,carries out weighted processing and word frequency statistics of feature words according to the source location of feature words,constructs a book feature word sparse matrix,carries out naive Bayes calculation on the book feature word sparse matrix with book classification number proportionally,finds the maximum probability of book classification,and evaluates and trains the classification model.The experimental results show that the book intelligent classification model based on weighted selection of feature words using naive Bayes algorithm has good practicality,it is effective and feasible for further improving the intelligence and efficiency of the collection and editing department’s work.

关 键 词：机器学习 朴素贝叶斯 图书智能分类 中文图书 

分 类 号：G254.3[文化科学—图书馆学]