Understanding client satisfaction of prefabricated curtain wall in Hong Kong using XGBoost and Pearson correlation

Abstract

Purpose

The unitized curtain wall system (UCWS), a symbol of modern architecture, is gaining popularity among prefabricated components. Previous studies have focused on both construction technology advances and material selection strategies to facilitate the UCWS. However, the topic of client satisfaction, which drives industry development by targeting clients' demands, has gone unnoticed. Therefore, the current study aims to investigate client satisfaction with UCWS products in Hong Kong by finding its influential factors.

Design/methodology/approach

A systematic review was employed to first identify the influential factors. A semi-structured interview was employed to validate the reliability of the extracted factors. The machine learning algorithm Extreme Gradient Boosting (XGBoost) and the Pearson correlation were then employed to rank the importance and correlation of factors based on the 1–5 Likert scale scores obtained through a questionnaire survey.

Findings

The findings revealed that “reduction in construction time” and “reduction in construction waste” are the most important factors and have a strong positive influence on client satisfaction.

Originality/value

Unlike previous studies, the present study focused on a novel research topic and introduces an objective analysis process using machine learning algorithms. The findings contribute to narrowing the knowledge gap regarding client preference for UCWS products from both individual and collaborative perspectives, providing decision-makers with an objective, quantitative and thorough reference before making investments in the curtain wall management development.

Citation

Kwok, T.W., Chang, S. and Li, H. (2023), "Understanding client satisfaction of prefabricated curtain wall in Hong Kong using XGBoost and Pearson correlation", Engineering, Construction and Architectural Management, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/ECAM-03-2023-0276