Hybrid Machine Learning Approach Enhances Sustainability in Green Buildings

Strategic Insights -

Despite the growing adoption of green buildings, robust and data-driven approaches for assessing and predicting project risks remain limited. Research from the East China University of Science and Technology proposed an innovative hybrid framework combining the fuzzy analytic hierarchy process (FAHP), multilayer perceptron neural networks (MLPNNs), and particle swarm optimization (PSO) to quantify and forecast the impact of critical risks on green buildings' performance. The study highlighted the importance of economic, market, and functional risks and demonstrated strong predictive capability through optimized framework results. The findings provide actionable, quantitative risk predictions under uncertainty, substantiating the framework's effectiveness in supporting sustainability.

Key Takeaways:

  • Researchers from the East China University of Science and Technology developed a hybrid machine learning framework to assess and predict project risks in green buildings.
  • The framework combines FAHP, MLPNNs, and PSO to quantify and forecast the impact of critical risks, including economic, market, and functional risks.
  • The study identified ten risk categories and prioritized them based on structured input from 30 domain experts in Shenzhen, China.
  • The optimized framework achieved RMSE values ranging from 0.06 to 0.09 and R2 values of up to 0.95 across all outputs, demonstrating strong predictive capability.
  • The research emphasized the importance of addressing critical risks in green building development to ensure sustainability and performance.
  • The findings of the study provide a data-driven approach for assessing and predicting project risks, supporting the growth of sustainable development in the construction industry.
  • The researchers involved in the study included Yanqiu Zhu, Hongan Chen, Jun Ma, and Fei Pan.

Statistics:

  • 30 domain experts in Shenzhen, China, contributed to the study by providing structured input on risk categories.
  • 10 risk categories were identified and prioritized based on expert input.
  • Economic, market, and functional risks emerged as the most influential risk categories.
  • The RMSE values of the optimized framework ranged from 0.06 to 0.09.
  • R2 values of up to 0.95 were achieved across all outputs, demonstrating strong predictive capability.

Sources:

  • Risk Management of Green Building Development: an Application of a Hybrid Machine Learning Approach Towards Sustainability. Sustainability, 2025;17(14):6373.
  • NewsRx LLC. Investigators from East China University of Science and Technology Release New Data on Sustainability Research. Ecology, Environment & Conservation. August 29, 2025; p 195.