Predicting Construction Quality Factors Using Artificial Neural Networks: An Analysis of Key Influences in Birendranagar, Surkhet, Nepal

Abstract

This study examines the key factors influencing construction quality in Birendranagar, Surkhet, focusing on manpower, equipment, management, funding, and duration-related challenges. A questionnaire survey was conducted with 59 respondents, and data were analyzed using the Relative Importance Index (RII) and Mean Value Response (MVR). The findings reveal that insufficient worker training (RII = 0.78), weak site supervision (RII = 0.81), outdated machinery with poor maintenance (RII = 0.82), and financial constraints (RII = 0.84) significantly impact construction quality. Additionally, poor planning and scheduling (RII = 0.83) contribute to delays and inefficiencies.

An Artificial Neural Network (ANN) model using IBM SPSS 27 was employed to predict the most influential factors affecting construction quality. While RII rankings identified funding-related factors as the most critical, ANN analysis indicated that machine efficiency, management, and project duration have a greater impact, highlighting a gap between stakeholder perceptions and data-driven insights. Sensitivity analysis further revealed that machine-related factors hold the highest importance (100%), followed by management (96%) and duration (85%), while manpower factors ranked lowest (68%).

The study emphasizes the need for workforce training, modern equipment maintenance, financial stability, and improved site management to enhance construction quality. A balanced approach that integrates perception-based insights with data-driven findings is recommended for effective decision-making and sustainable improvements in construction projects. Future research should explore labor union impacts, cultural influences, client involvement, regulatory compliance, and risk mitigation strategies to further enhance construction quality management