AN INNOVATIVE HYBRID MACHINE LEARNING TECHNIQUES FOR PREDICTING CONSTRUCTION COST ESTIMATES

Abstract

The importance of precise cost estimations in construction investment sector cannot be overstated, serving as the cornerstone for crucial investment decisions. Unfortunately, inaccurate estimations often lead to significant losses due to flawed investment objectives. Consequently, the pursuit of an effective investment decision support system has become a paramount research focus, aiming to aid construction investors in making informed and timely choices. This study focuses on a novel hybrid machine learning (HML) approach, amalgamating various base models such as artificial neural networks (ANNs), support vector machines (SVMs), multiple linear regression (MLR), decision trees (DTs) and random forest (RF) to construct a sophisticated construction cost forecasting model. Remarkably, empirical findings demonstrate the exceptional accuracy of the hybrid ANN-DT model, reaching an impressive 92.1% and surpassing individual models. This breakthrough holds the promise of substantial advantages and increased profitability within the construction industry, particularly benefitting professionals in civil engineering, architecture, and construction investing. By combining the predictive strength of ANNs with the transparent decision rules of DTs, this hybrid model effectively addresses the industry's need for precise predictions and understandable forecasting methodologies, representing a significant advancement in enhancing informed decision-making processes within the construction domain.

Keywords

• construction cost estimation
• machine learning
• hybrid learning model
• residential buildings