Financial Distress Early Warning for Real Estate Listed Companies of China Based on Elliptical Space Probabilistic Neural Networks

In order to identify the financial risk for real estate listed companies of China, this paper uses Elliptical Space Probabilistic Neural Networks (ESPNN), which is the improvement of the probabilistic neural network. It has three network parameters: (1) variable weights representing the importance of input variables; (2) the reciprocal of kernel radius representing the effective range of data; and (3) and data weights representing the data reliability. The paper adopts ESPNN with annual financial statement data of real estate listed companies from the A-share market of China based on 18 financial and non-financial indicators. The results showed that the average accuracy of the financial distress early warning model reaches up to 95% above. Then, we compared ESPNN with a support vector machine (SVM) and Probabilistic Neural Networks (PNN), showing that ESPNN is slightly more accurate than PNN and much more accurate than SVM. In the context of deepening adjustments in the Chinese real estate industry, this model not only provides reference indicators for business managers and market investors, but also helps policy makers evaluate the potential risks in the real estate industry in a timely fashion.