Analysis and Optimization of Housing Construction Financing Methods Using Artificial Neural Networks

Abstract

Financing housing construction projects has always been considered one of the most significant challenges in the construction industry. Selecting an appropriate financing method has a direct impact on project profitability, execution time, and overall project success. The complexity of the relationships among economic, technical, and managerial factors has reduced the efficiency and accuracy of traditional decision-making approaches in many cases. Therefore, this study employs Artificial Neural Networks (ANNs), as one of the modern Artificial Intelligence (AI) tools, to optimize the selection of financing methods in housing construction projects. In this research, three types of neural networks, including the Multi-Layer Perceptron (MLP), Neuro-Fuzzy Network, and Radial Basis Function (RBF) Network, were designed and modeled. Influential parameters related to financing and construction costs were selected as input variables, while the final project profit was considered the output variable of the models. To determine the optimal structure, various models with different architectures and functions were trained and evaluated using error indices. Furthermore, sensitivity analysis was conducted on the selected models to assess the impact of each input variable on model performance. The results revealed that the MLP network outperformed the other models in terms of prediction accuracy and optimization capability for housing financing methods. In addition, the sensitivity analysis indicated that the pre-sale parameter had the greatest influence on the final profit and financial success of construction projects. The findings of this study demonstrate that ANN can serve as an intelligent and efficient tool for improving financial decision-making processes in housing construction projects and reducing project-related economic risks.