Intelligent Identification of Corporate Tax Evasion Based on LM Neural Network
作者单位：School of Information and Control EngineeringShenyang Jianzhu University School of Municipal and Environmental Engineering Shenyang Jianzhu University School of ScienceShenyang Jianzhu University
会议名称：《第37届中国控制会议》
会议日期：2018年
学科分类：0202[经济学-应用经济学] 02[经济学] 020205[经济学-产业经济学] 020203[经济学-财政学（含∶税收学）]
基 金：supported by Shenyang Jianzhu University Youth Foundation under Grant 2015032 supported by Liaoning Province Science Foundation under Grant LJ2016004
关 键 词：Tax evasion automobile sales enterprises LM neural network
摘 要：Automobile sales enterprises are an important part of tax revenue in China. Tax evasion is not conducive to the healthy development of the market. It has seriously damaged the national interest and damaged the fairness and justice of the automobile sales industry. This paper proposes an intelligent identification model based on big data for tax evasion in automobile sales enterprises. Using existing reports, selecting the maintenance margin, the VAT tax burden, agent insurance and license income as indicators, disquieting data and construct an expert sample, training LM neural network model. Results show that: The recognition accuracy of the model was 89%, and the results were assessed by ROC curve, which showed that the classification effect was good.