Research On The Model Of Audit Opinoin Prediction Based On Grey System Theory And Neural Network

The disclosed financial reports of listed company are the basis of decision making by the stakeholders.As an independent third party,the CPA shall issue an audit opinion on the financial report of the audited entity.The audited financial report plays a very important role in enhancing the credibility of financial information.The audited financial statements are more reliable,so that information users can be more confident,and can reduce investment decision-making errors.The type of audit opinion issued by certified public accountants has a significant impact on the investment decision behavior by the stakeholders.Therefore,it has certain practical significance to research on the prediction of audit opinion types of listed companies.This paper constructs the audit opinion prediction model based on grey system theory and neural network.In order to ensure the prediction effectiveness,financial indicators are selected based on previous research,and non-financial indicators are also selected which include two aspects about corporate governance and accounting firm.This paper selects manufacturing companies from Shanghai and Shenzhen A-share markets as the research sample in 2013-2015,which are issued non-standard audit opinion.According to the company’s assets and the same industry,it chooses an equal number of sample which are issued standard unqualified audit opinion.And after removing the missing values,the training group,the test group and the prediction group are finally obtained.and neighborhood rough set method is used to screen training data and retain the important indicators to reduce the complexity of model training.The training sample is used in model training and simulation based on BP neural network.The test group sample is used to test the established model.The sample data of the prediction group is selected to construct the audit opinion prediction model based on grey system theory and neural network.This model combines the advantages of the grey system theory with less information to construct the model and neural network’s adaptability.The empirical results show that the prediction accuracy of the comprehensive model constructed with both financial and non-financial indicators is higher than the model based only on financial indicators.The audit opinion prediction model based on grey system theory and neural network shows good prediction ability.