Research On Financial Crisis Early Warning Of Real Estate Listed Companies Based On Multi-classifier

As a macroeconomic pillar industry,the real estate industry has received fierce control over macroeconomic policies in recent years.Its commodity prices have undergone severe volatility.Real estate listed companies,as commodity suppliers to the entire market,have directly determined the real estate's financial status.Market development is healthy or not.Therefore,this article will be a real estate listed company as a window of the property market to conduct research,through the establishment of real estate listed companies financial early warning model,the real estate company's financial status of early warning,no doubt to guard against risks,resolve potential debt crisis,the real economy deleveraging are important The practical guiding significance.Through the study of relevant literature,scholars often fail to analyze the real estate company's operating characteristics and financial characteristics when building real-estate companies' early-warning index systems.The selected financial and non-financial indicators often fail to reflect the characteristics of the real estate industry.When modeling,often the warning results are not ideal.Due to the long development cycle of real estate companies,the sales model is mainly the pre-sale system,revenue recognition and cash flow do not match,current operating income often reflects the operating results of 1-2 years ago,the current sales receipt can only be included in the advance receipt.In the accounts,resulting in accounting information to some extent there is lag and non-reality.The non-financial data of housing enterprises' land reserves can effectively measure the size and strength of enterprises.Now the state strictly controls the planning and use of land use,and at this time,the land reserves can greatly affect the future years.In the construction of the project,the larger the land reserve of the housing enterprises,the more real estates will be developed in the future,and the future performance will also grow faster.The land area and amount will be the most advanced indicators in the forward-looking information of the housing enterprises,which can best reflect The company's future revenue and profit levels.Therefore,in the construction of the real-estate financial warning index system,this paper fully considers the characteristics of the above-mentioned real estate industry,and restructures traditional financial indicators from the five dimensions of profitability,operational capacity,capital structure,debt repayment ability,and growth ability,and introduces “net cash”."Return on assets," "real assets and liabilities," "cash turnover," "cash inventory turnover," and "earnings per share," and other financial and non-financial indicators.Strive to be able to reflect the business and financial characteristics of real estate companies in all aspects and from multiple perspectives,thereby greatly improving the accuracy of financial early warning.In the empirical model,this paper differs from the single model in early warning research in that it compares multiple combinations of optimized classifiers in data mining with traditional logit models,and establishes a multi-dimensional evaluation system to evaluate empirical results.Strive to build a scientific and reasonable early warning model.The final empirical results show that: 1.Most of the early warning indicators with industry characteristics constructed in this paper have passed the significant level test,while the traditional financial indicators,such as "current ratio" and "quick move ratio" have not passed the significance level test.This shows that the indicators constructed in this paper have a significant role in distinguishing whether a real estate company is in financial crisis.2.The accuracy of data mining algorithms in both training set and forecasting set far exceeds the logit model,especially the precision of the forecasting set.Because of the different data set partitioning,there are large fluctuations in the prediction accuracy.The company,the classification accuracy and stability of data mining algorithms are higher than the logit model.Through the analysis of multi-dimensional evaluation indicators,the PSO-SVM model was finally selected as a financial early-warning model for real estate listed companies.