摘要
针对近年来图书馆读者流失严重的问题,作者从读者和图书馆两方面因素进行分析,将这些文本类信息纳入到卷积神经网络,搭建图书馆读者流失分析模型(TEXT-CNN)。以西安市某图书馆为例,对读者流失进行预测并生成读者列表,将该模型所用算法与目前常用的预测模型(如遗传算法优化的BP神经网络GA-BP模型、支持向量机SVM模型以及逻辑回归Logistic模型)进行比较。结果显示,这里提出的TEXT-CNN模型的预测误差最小,精确率和准确率最高,同时F1值和召回率也优于其余三种模型,这表明将所提出的模型用于对图书馆读者流失进行预测是可行的。本研究对于图书馆提前预知读者流失情况并及时采取预防措施具有现实指导意义。
In view of the serious problem of reader loss in libraries in recent years,this article analyzes the two factors of readers and libraries,incorporates these textual information into the convolutional neural network,and builds a library reader loss analysis model(TEXT-CNN).Taking a library in Xi’an as an example,it predicts the loss of readers and generates a reader list,and the algorithm used in the model is compared with the currently commonly used prediction models(such as genetic algorithm optimized BP neural network GA-BP model,support vector machine SVM model and logistic regression Logistic model).The results show that the TEXT-CNN model proposed in this paper has the smallest prediction error,the highest precision and accuracy,and the F1 value and recall rate are also better than the other three models.This shows that the model proposed in this paper is feasible to predict the loss of library readers.This research has practical guiding significance for the library to predict the loss of readers in advance and take preventive measures in time.
作者
徐根祺
曹宁
谢国坤
董三锋
张正勃
XU Genqi;CAO Ning;XIE Guokun;DONG Sanfeng;ZHANG Zhengbo(Mechanical Electric Engineering Department,Xi’an Traffic Engineering Institute,Xi’an 710300,China;Civil Engineering Department,Xi’an Traffic Engineering Institute,Xi’an 710300,China)
出处
《微型电脑应用》
2022年第9期5-7,15,共4页
Microcomputer Applications
基金
西安交通工程学院中青年基金项目(2022KY-48)
西安交通工程学院中青年基金项目(2022KY-33)
陕西省教育厅专项科学研究计划项目(17JK1019)。
关键词
图书馆
读者流失
文本信息
libraries
reader loss
text information
