A Study Of Combination Forecasting Model On Loss Of Assets For State﹣Owned Listed Companies In China

When China's reform and opening up is gradually in depth, state‐owned enterprises continue to grow, but at the same time, the loss of state assets has intensified, so finding prompt solution has become a major problem for government departments. The situation of huge loss of state﹣owned enterprise assets is not easy, it is a gradual, incremental process, so the loss of state assets should have the characteristics of aura, that is, the loss of state assets can be predicted. In fact, ifwe can accurately predict the loss of state﹣owned enterprise assets, it will be of great significance.The purpose of this paper is to analyze the process of state enterprises and the reasons for the problems, and explore the method of accurately predicting the loss of assets for state﹣owned listed companies.Based on the above objectives, firstly we review relevant literature on the loss of assets. Secondly we find the detailed definition of loss of state assets. Then we introduce the MDA, Logistic regression analysis, artificial neural networks and support vector machine model, and on the basis, building a forecasting combination model.We select 25 variables as the core indicators to predict the loss of state assets. On the basis, the statistical methods select seven indicators as input variables for forecasting model. The results are that:among the individual forecasting model, the predicting accuracy is different, but because of the combination forecasting model is a collection of statistical models and artificial intelligence models, the predicting accuracy is higher than the four single forecasting models, which has a strong advantage and value.Through this study, we can draw the following conclusions:(1)The loss of State assets can be predicted by some of the financial indicators and non﹣financial indicators.(2)The four single forecasting model can accurately predict the loss of state﹣owned assets of listed companies, but, comparing to the statistical models, artificial intelligence models have higher predicting accuracy.(3)Combination forecasting model can accurately predict the loss of state﹣owned assets of listed companies, and its predicting accuracy is not less than the four single forecasting model.Finally, based on the theory of principal﹣agent theory, we actively explore the measures to prevent the loss of state‐owned assets.