基于随机森林的财务报表隐私数据自动加密方法研究  

Research on Automatic Encryption Methodof Financial Statement Privacy Data Based on Random Forest

作　　者：晏银芳 田维维 YAN Yin-fang;TIAN Wei-wei(School of Economics and Management,Huizhou Economics and Polytechnic College,Huizhou 516057,Guangdong,China;School of Accounting and Finance,Anhui Institute of International Business,Hefei 230031,China)

机构地区：[1]惠州经济职业技术学院经济管理学院,广东惠州516057 [2]安徽国际商务职业学院财会金融学院,安徽合肥230031

出　　处：《兰州文理学院学报（自然科学版）》2023年第4期46-51,共6页Journal of Lanzhou University of Arts and Science(Natural Sciences)

基　　金：安徽省教育厅优秀青年人才支持计划重点项目(gxyqZD2020148)。

摘　　要：为保证财务报表隐私数据的安全性,避免在攻击时发生数据泄露,本文对基于随机森林的财务报表隐私数据自动加密方法进行了研究.通过多个回归树和决策树构建随机森林回归模型,依据该模型分类财务报表数据中的隐私和非隐私数据,基于复合混沌序列系统生成报表隐私数据的混沌随机序列,同时依据二值化寻优方法完成序列编码;通过密钥控制方法,设计财务报表隐私数据加密密钥后,利用特征聚类和编码融合自动加密财务报表隐私数据.测试结果显示:该方法能较好完成隐私数据和非隐私数据的分类,具备良好的随机性,Lyapunov指数均在0.52以上;加密后数据之间的相关系数结果均在0.1以下,面临不同的攻击时,用户响应值均在0.85以上,保证财务报表隐私数据的安全,避免数据泄露.In order to ensure the security of financial statement privacy data and avoid data leakage during attack,this paper studies the automatic encryption method of financial statement privacy data based on random forest.The random forest regression model is constructed through multiple regression trees and decision trees.According to the model,the private and non-private data in the financial statement data are classified.The chaotic random sequence of the report privacy data is generated based on the composite chaotic sequence system,and the sequence coding is completed according to the binary optimization method;through the key control method,after designing the private data encryption key of financial statements,the private data of financial statements is automatically encrypted by feature clustering and coding fusion.The test results show that this method can reliably complete the classification of private data and non-private data,and the classification effect is good.It has good randomness,and the results of Lyapunov exponent are above 0.52.The correlation coefficient results between the encrypted data are below 0.1.In the face of different attacks,the user response value results are above 0.85,so as to ensure the security of the private data of the financial statements and avoid the risk of data disclosure.

关 键 词：随机森林 财务报表 隐私数据 自动加密 复合混沌序列 序列编码 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]