The Research On Business Performance Evaluation Of Listed Real Estate Companies Based On SVM Regression

To evaluate the operation performance of listed real estate companies is an inevitable requirement for these companies to deal with the challenge of stock elimination. Which is also an essential part for real estate companies to survive in the increasingly fierce market competition. Evaluating the performance of their own business, enterprises recognize their own strengths and weaknesses, which is useful for them to improve management standard and increase their own competitiveness. Currently the support vector machine method is a hot academic research, especially its using in classification and regression is recognized by more and more people. Applying the support vector machine regression method to the performance evaluation of listed real estate companies is an combination of advantage of SVR algorithm and the practical needs of real estate companies performance evaluation.The evaluation index selection principles was designed on the basis of analyzing the characteristics of listed real estate companies, as well as the ideological basis of support vector machine theory. Ten indicators was selected from numerous financial indicators to composite the index system, and these indicators reflect four areas of the companies business performance which includes investment and income, operating capabilities, solvency and capital structure.In terms of model building and examples demonstrate,96 A-share listed real estate companies financial data was extracted from Shanghai and Shenzhen stock exchange market. In order to get the value of comprehensive evaluation TOPSIS method was used, as well as the entropy method to calculate the weight of each property on the basis of data assimilation and standardization, with the SVR model being built. The model’s fitting effect which is the output variables of test samples was test by comparing with the fitting result of RBF neural network. The results showed that the model based on support vector machine regression fitting effect is better than that of RBF neural network. Therefore in terms of listed real estate companies performance evaluation, the SVR model shows stronger predictive power.Through analyzing of listed real estate companies operating performance concluded that the performance of listed real estate companies in our country was uneven, and the overall pefrmence was low, well between the considerable strength of the enterprise competition is more intense. Among all the indicators impacting the performance of listed real estate company, the earning per share index maximum weighted. Therefore, when we conduct the research of listed companies performance, we should focus more attention on the earning per share index, but it is not sufficient to replace the role and function of other indicators.I hope that these research results will have some reference value for the listed real estate companies conducting performance evaluation.