基于蝙蝠算法优化LSSVM的图书馆访问流量预测

Library access traffic prediction based on bat algorithm optimized for LSSVM

为提高图书馆访问流量预测的精度,对于LSSVM预测效果易受惩罚参数γ和径向基核函数宽度g的影响内容进行相应的分析研究,在此基础上对惩罚参数γ和径向基核函数宽度g进行优化,提出一种基于蝙蝠算法优化LSSVM的图书馆访问流量预测算法。与LSSVM和BP相比较,能够得出如下结论:通过BA算法优化之后所得出的LSSVM能够加大程度的增强图书馆访问流量预测的准确度,通过全新算法的实施可以使得图书馆访问流量预测更加精准,从而针对性的提升服务质量。

In order to improve the accuracy of library traffic prediction, the impact of the LSSVMprediction effect on the penalty parameters γ and the radial basis kernel function width g is analyzed,andthe penalty parameters γ and the radial basis kernel function width g is optimazed, and then libraryaccess traffic prediction algorithm is proposed based on BA-LSSVM. Compared with LSSVM and BP, theresults show that BA-LSSVM can effectively improve the accuracy of library access traffic prediction andprovide new methods and approaches for library access traffic prediction.