Scholarlink

<p>This study addresses the growing need for sustainable construction by systematically reviewing the integration of machine learning (ML) techniques into life cycle assessment (LCA) practices within the construction sector. While previous reviews have addressed ML or LCA separately, this paper uniquely maps specific ML algorithms including random forest, extreme gradient boosting, and artificial neural networks to their applications in environmental prediction, carbon emission modelling, material optimization, and sustainability assessment. A targeted literature search yielded 21 relevant studies spanning buildings, materials, infrastructure and energy systems. The review highlights that regression models dominate current ML–LCA integration due to their high accuracy and adaptability, while clustering and classification techniques remain underutilized. Ensemble and hybrid approaches, particularly stacking and deep learning variants, show superior performance in improving robustness and predictive power. Through bibliographic and thematic analyses, the study also identifies leading contributors, emerging trends, and research gaps such as limited standardization, data quality inconsistencies and high computational costs. The final goal of this review is to guide future research by outlining the performance trends of various ML models, emphasizing the need for interpretability, and identifying opportunities for integration with tools like building information modelling. By establishing a comparative benchmark of ML applications in LCA, this review provides a foundation for more consistent, transparent, and impactful use of data-driven methods in sustainable construction.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Advances in integration of machine learning and life cycle assessment within construction sector: a literature review

  • 期刊论文

  • 发表日期：

    2025年6月26日

查看全文
  • Samrity Jalota,
  • Mohammad Farhad Ayazi

摘要

This study addresses the growing need for sustainable construction by systematically reviewing the integration of machine learning (ML) techniques into life cycle assessment (LCA) practices within the construction sector. While previous reviews have addressed ML or LCA separately, this paper uniquely maps specific ML algorithms including random forest, extreme gradient boosting, and artificial neural networks to their applications in environmental prediction, carbon emission modelling, material optimization, and sustainability assessment. A targeted literature search yielded 21 relevant studies spanning buildings, materials, infrastructure and energy systems. The review highlights that regression models dominate current ML–LCA integration due to their high accuracy and adaptability, while clustering and classification techniques remain underutilized. Ensemble and hybrid approaches, particularly stacking and deep learning variants, show superior performance in improving robustness and predictive power. Through bibliographic and thematic analyses, the study also identifies leading contributors, emerging trends, and research gaps such as limited standardization, data quality inconsistencies and high computational costs. The final goal of this review is to guide future research by outlining the performance trends of various ML models, emphasizing the need for interpretability, and identifying opportunities for integration with tools like building information modelling. By establishing a comparative benchmark of ML applications in LCA, this review provides a foundation for more consistent, transparent, and impactful use of data-driven methods in sustainable construction.