Intelligent Identification of Corporate Tax Evasion Based on LM Neural Network

Automobile sales eriterprises 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.