基于小波支持向量机的图书馆借阅量预测  

Prediction of Library Loan Volume Based on Wavelet Support Vector Machine

作　　者：余梦媛 YU Mengyuan(Department of Library,Henan Polytechnic Institute,Nanyang,Henan 473000,China)

机构地区：[1]河南工业职业技术学院图书馆,河南南阳473000

出　　处：《微型电脑应用》2020年第4期150-152,共3页Microcomputer Applications

摘　　要：当前图书馆借阅量预测方法无法描述混沌性等非平稳变化特征,导致图书馆借阅量预测错误差大,为了改善图书馆借阅量的预测效果,设计了基于小波支持向量机的图书馆借阅量预测方法。首先对当前国内外图书馆借阅量的预测研究现状进行分析,找到引起图书馆借阅预测误差大的原因,然后收集图书馆借阅量预测的历史数据,并通过混沌分析算法对历史数据进行重新构造,并引入小波支持向量机实现图书馆借阅量预测模型的建立,最后与其它图书馆借阅量的预测方法在相同环境进行对比测试。提出的方法可以对图书馆借阅量的变化特征进行深度挖掘,图书馆借阅量预测精度超过95%,高于对比方法图书馆借阅量预测精度,获得更加可靠的图书馆借阅量的建模和预测结果。Current library borrowing forecasting methods cannot describe the non-stationary change characteristics such as chaos, which leads to great error in library borrowing forecasting. In order to improve the forecasting effect of library borrowing, a library borrowing forecasting method based on wavelet support vector machine(WSVM) is designed. Firstly, this paper analyzes the current situation of the prediction of library borrowing volume at home and abroad, finds out the reasons for the large error in the prediction of library borrowing volume, then collects the historical data of library borrowing volume prediction, reconstructs the historical data through chaotic analysis algorithm, and introduces the wavelet support vector machine to establish the prediction model of library borrowing volume. Finally, it establishes the prediction model of library borrowing volume. The forecasting methods of library borrowing volume are tested in the same environment. The presented method can deeply mine the changing characteristics of library borrowing volume. The prediction accuracy of library borrowing volume exceeds 95%, which is higher than that of the comparative methods. It can obtain more reliable model and prediction results of library borrowing volume.

关 键 词：图书借阅量 混沌性特征 数据挖掘 混沌分析算法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]