Material Requirement Prediction System and 2D Visualization of Light Steel Roof Design Using the Decision Tree Algorithm Based on an Expert System

Abstract

Material requirement planning is a critical aspect of the construction industry, which is increasingly moving toward digital transformation. However, in practice, material estimation for roof construction is still predominantly performed using conventional manual calculations. This approach relies heavily on subjective judgment, resulting in inconsistent estimates and increasing the risk of cost overruns, material waste, and project delays. To address these challenges, this study proposes an expert system-based web application for predicting material requirements and generating two-dimensional (2D) visualizations of light steel roof designs The proposed system employs the Classification and Regression Tree (CART) algorithm, a Decision Tree technique, to classify roof specifications based on building dimensions, while an expert system provides rule-based recommendations for supporting construction materials, including roof sheets, screws, and structural components. Users are only required to input the building length and width, after which the system automatically predicts material quantities, estimates construction costs, and generates a 2D roof design visualization. The integration of predictive analytics and expert knowledge minimizes human error, improves the consistency of material estimation, and enhances decision-making for both technical and non-technical users. Furthermore, the developed system provides a transparent, efficient, and user-friendly solution for construction planning by reducing calculation complexity and optimizing material utilization. The proposed approach demonstrates the potential.