Research On The IDSS Modeling Of Real Estate Credit Risk Based On Complex Multi-projects

With such rapid development of real estate, as a result, that means the rise of industryconcentration and the sharp increasing in the number of items of real estatecompanies.Therefore, complex multi-project management model has become thenorm in the form of real estate business management. This model brings greatereconomies of scale to the real estate business, however, there are more diverse, morecomplex risk factors, making the Bank more complex and difficult in its credit riskassessment. This thesis aims to build a real estate credit risk IDSS based on complexmulti-project, using a variety of smart technologies and methods of artificialintelligence, Such as the Multi-agent technology, artificial neural networks and expertsystems, making the intelligent decision support system more intelligent, flexible, atthe same time, getting a certain theoretical significance and research value.The main research work in this thesis focused on the following aspects:(1) To recognize the main risks of complex multi-project real estate enterprises, andobtained the risk factor index system.(2) To build a whole framework of intelligent decision support system based onMulti-agent risk assessment, and the detailed design.(3) To build the model library embedded in the IDSS risk assessment model designand achieve it.Through the analysis and research, the following findings and conclusions have beenachieved:(1) After Understanding the real estate multi-project development companies throughaccessing extensive literature, and identifying the risks based on complexmulti-project real estate enterprises, the main risks have been obtained by twocategories: One is the risk of corporate headquarters, including the Industry risk,internal management risk and environmental risk; Another is the risk of the projectdevelopment, such as various risks encountered in the course of the projectdevelopment cycle, and various types of risk are carried out a detailed analysis. (2) Multi-agent technology is integrated into the intelligent decision support systems,and a overall framework based on the IDSS of multi-agent real estate credit risk isconstructed to solve the control problem for the credit risk of real estate companies, atthe same time, conduct a detailed functional and structural design of the main agenttypes and the database, model base, method base, Knowledge Base in the systemframework respectively. Finally, analyze the communication process between agentsin the system framework. Through the introduction of the Multi-agent technologymakes the real estate credit risk IDSS in this thesis having the features of the moreintelligent, more flexible.(3) Studying based on the build of IDSS embedded in the model library based on theMulti-agent risk assessment model, this thesis is the integration of risk assessmentmethods in neural networks and expert systems, which makes the intelligent decisionsupport system synthesize multi-agent technology, neural network and expert systemtechnology. This system has a more powerful intelligence, provides a more reliabletool for risk assessment.(4)With an example analysis, through understanding and studying the multi-projectmanagement in a real estate enterprise in Qingdao, an index system concerning theenterprise risk factor has been established. BP neural network training and testingsamples of data are drawn by the expert survey score, and achieving the BP network’straining and learning and the save of weights and threshold values by using MATLAB,and achieve the integration of neural networks and expert systems, through detectingthe last two projects of the enterprise, there is a tiny error between the network outputand the experts’ result. Therefore, in this thesis, the design of multi-intelligenttechnology-based IDSS has a profound practical significance of the real estate riskassessment.