基于TSO-Elman模型的财务困境预警  

Financial Distress Early Warning Based on TSO-Elman Model

作　　者：康亮河 赵正强 万君 王剑恩 王博[1] KANG Lianghe;ZHAO Zhengqiang;WAN Jun;WANG Jianen;WANG Bo(Gansu Agricultural University,Lanzhou 730070,China)

机构地区：[1]甘肃农业大学,甘肃兰州730070

出　　处：《现代信息科技》2023年第9期162-165,共4页Modern Information Technology

基　　金：2022年甘肃农业大学大学生创新创业训练计划项目(202216017)。

摘　　要：针对Elman神经网络在财务困境中预测精度低的问题,文章提出了结合金枪鱼TSO算法与Elman神经网络的财务困境预警模型。针对2021年A股上市公司的财务数据,首先采用Z分数及多重插补法对异常值及缺失值进行了处理,并利用PCA算法筛选出贡献度为90%的重要属性;其次,利用TSO算法对Elman神经网络参数寻优,构建财务困境预测模型。实验结果表明:TSO-Elman模型在对上市公司财务困境预测中具有较高的预测精度及稳定性,相比RBF、BP、SVM及Elman模型,其分类准确率分别提高了18.89%、14.13%、12.54%及4.61%。Aiming at the problem of low prediction accuracy of Elman neural network infinancial distress,this paper proposes afinancial distress early warning model combining Tuna Swarm Optimization(TSO)algorithm and Elman neural network.In view of thefinancial data of A-share listed companies in 2021,the Z score and multiple imputation method are used to process the outliers and missing valuesfirstly,and the PCA algorithm is used to screen out the important attributes with a contribution of 90%.Secondly,TSO algorithm is used to optimize Elman neural network parameters to construct afinancial distress prediction model.The experimental results show that the TSO-Elman model has high prediction accuracy and stability infinancial distress prediction of listed companies.Compared with RBF,BP,SVM and Elman models,its classification accuracy has improved 18.89%,14.13%,12.54%and 4.61%respectively.

关 键 词：TSO算法 ELMAN神经网络 PCA算法 财务困境 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP183[自动化与计算机技术—计算机科学与技术]