Artificial Neural Networks Improve Green Building Project Success Rates

Strategic Insights -

Research by the National Taiwan University of Science and Technology, funded by the Taiwan Building Technology Center and the Ministry of Education in Taiwan, has demonstrated the effectiveness of artificial neural networks in predicting the success of green building projects. By analyzing data from 354 green building projects in Taiwan, the study found that different types and grades of green buildings are based on varying green building technologies. The researchers constructed a prediction model using an artificial neural network that was able to accurately predict green building grades and costs, achieving an accuracy rate of over 80%.

Key Takeaways:

  • The study conducted two-stage data mining on 354 green building projects in Taiwan to identify issues in the preliminary design phase, including technology adoptions, green building grades, and construction costs.
  • The researchers found that different types and grades of green buildings are based on varying green building technologies, with a high-grade green building placing more emphasis on green building technologies such as air conditioning, CO2 reduction, and indoor environments.
  • The artificial neural network prediction model was able to accurately predict green building grades and costs, achieving an accuracy rate of over 80%.
  • The systematic data mining method developed in this study can assist architects and building owners in reducing preliminary design time and costs, as well as improving the success rates of green building projects.
  • The study concluded that the proposed approach can be adjusted for other regions with different climates or green building rating tools to construct more suitable applications.
  • The findings of this study have been peer-reviewed and published in the Journal of Building Engineering.

Statistics:

  • 354 green building projects in Taiwan were analyzed in the study.
  • The accuracy rate of the artificial neural network prediction model was over 80%.
  • The study found that green building technologies are affected by building types, regulations, costs, climate conditions, and geographic restrictions.
  • The study concluded that the systematic data mining method can effectively assist architects and building owners in reducing preliminary design time and costs.
  • The proposed approach can be adjusted for other regions with different climates or green building rating tools to construct more suitable applications.

Sources:

  • "Applying Data Mining Techniques To Explore Technology Adoptions, Grades and Costs of Green Building Projects" (Journal of Building Engineering, 2022;45:103669)
  • NewsRx. New Artificial Neural Networks Data Have Been Reported by Investigators at National Taiwan University of Science and Technology (Applying Data Mining Techniques To Explore Technology Adoptions, Grades and Costs of Green Building Projects). Network Weekly News. January 3, 2022; p 427.