Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Loan forecasting for commercial real estate with neuro-fuzzy techniques

Tsouvalakis Georgios

Full record

---

Summary

The buying and selling of real estate first appeared in the ancient years, the agreements which they always took place at the local level and did not go beyond the borders of the respective country. Over the years, investments in real estate intensified, deals had the option to be done with the help of third parties, and markets went beyond borders and became international. As global market integration progressed, globalization in the economy became an integral part of each economy. Cross-border capital flows increased asset purchases, and real estate markets played an important role in the market.It is noteworthy that one of the main advantages of real estate investments is that they provide a safety margin through annual cash flows, a fact that ensures the wealth of the individual owner, especially in periods of economic recession. Also, another advantage is that assets can be acquired at prices that offer attractive returns for a long period of time and in turn bring more profit to the investor. On the contrary, however, they can also bring about negative consequences if wrong actions are carried out and wrong decisions are taken respectively, with the result that they may even lead to bankruptcy. Real estate is divided into two groups: First, residential real estate and second, commercial real estate. This work focuses on the study of commercial real estate with the aim of analyzing forecasting models in order to capture the future image of commercial real estate loans, how it interacts and is affected by the global market. More specifically, commercial real estate is almost always comprised of an owner and a tenant while their purpose is commercial. The commercial real estate market is based on complex relationships of countless real estate markets and long-term factors. For this reason, the need for forecasts in the real estate market have found fertile ground and intense interest among researchers and professionals in order to avoid or minimize risks in order to make better decisions for the purpose of profit.In this thesis, a neuro-fuzzy technique is used which is based on the theory of neural networks, its implementation was done with the help of the MATLAB software. More specifically in the forecast of weekly commercial real estate loans (data dates from 06/02/2004 to 06/14/2023). The code fed the previous weeks' commercial real estate loan data, then applied it to the ANFIS models incorporating the fuzzy logic rules, and finally extracted the forecast value.In conclusion, the structure followed for the completion of the diploma work was: in the first chapters a literature review was done, then the basic methodologies and the theoretical background of the work necessary for the understanding of the study were formulated. In the last chapter, a more extensive analysis was made of the extracted conclusions that were analyzed with a mathematical background in the field of predictions and at the end the corresponding bibliography was recorded.

Available Files

Services