Predicting financial losses due to apartment construction accidents utilizing deep learning techniques

Area of Science:

• Construction Management
• Risk Assessment
• Data Science

Background:

• Increasing urban density drives rapid apartment construction.
• High-rise projects elevate accident frequency and financial loss severity.
• Climate change exacerbates construction site accident risks.

Purpose of the Study:

• To develop a quantitative financial loss prediction model for apartment construction accidents.
• To leverage deep learning for advanced risk management in construction.
• To support sustainable and efficient construction project management.

Main Methods:

• Collected and analyzed insurance claim payout data from South Korean construction sites.
• Applied deep learning algorithms to build predictive models.
• Focused on accidents occurring both inside and outside construction sites.

Main Results:

• A deep learning-based model for quantitative prediction of financial losses was generated.
• The model provides a framework for identifying and mitigating accident-related financial risks.
• The study demonstrates the efficacy of advanced algorithms in construction risk analysis.

Conclusions:

• The developed model is crucial for preventing and reducing financial losses in construction projects.
• Findings offer critical guidance for sustainable and effective construction project management.
• The framework can serve as a reference for future construction management research.