Machine Learning-based Feature Evaluation of the Factors Affecting the Effective Infrastructure Project Implementation
Year: 2025
Type: conference-paper
Abstract: Numerous factors affect how well infrastructure projects are implemented in different agencies and organizations, therefore identifying the important factors requires a data-driven approach. This study uses neural network simulations to examine the relative impact of various components using a feature evaluation based on machine learning. Neural network simulations yielded high predictive accuracy, with correlation coefficients (R) ranging from 0.97294 to 0.99084 and mean squared errors (MSE) between 0.014255 and 0.024731, validating the robustness of the model. Results indicate that the absence of a Project Procurement Management Plan (PPMP) and Annual Procurement Plan preparation is the most critical parameter in implementation requirements. In fund management, interagency funds not taken up in trust accounts emerged as the most influential factor, highlighting potential financial inefficiencies. Architectural & Engineering (A&E) and Administration & Supervision (A&S) expense allocations were found to be the primary determinants in the budget allocation category. For procurement of equipment, supplies, materials, and services, the absence of lease-to-own agreements significantly impacts cost-effective resource acquisition. Additionally, deviations from audit guidelines in rental fee disbursements pose financial accountability risks in the authority to charge fees for equipment rental. These results highlight the necessity of better financial supervision, procurement planning, and regulatory compliance to increase the efficiency and transparency of infrastructure projects, offering policymakers and agency administrators practical insights.
Language: English
Field: Decision Sciences
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W4416626348