Building loss assessment using deep learning algorithm from typhoon Rusa.

Area of Science:

• Climate Science
• Artificial Intelligence
• Disaster Management

Background:

• Climate crises, including extreme weather and natural disasters, cause significant global damage.
• Current efforts by the private sector and government to mitigate climate-related damage are insufficient to meet market demands.
• Typhoon-induced damage poses a substantial threat, necessitating advanced prediction methods.

Purpose of the Study:

• To develop and validate a deep learning algorithm for predicting typhoon damage.
• To provide a novel framework for enhancing disaster management and resilience against climate crises.
• To offer a tool for government agencies, facility managers, and insurance companies to predict and mitigate typhoon impacts.

Main Methods:

• Development of a Deep Neural Network (DNN) model.
• Model training using historical data from Typhoon Rusa damage.
• Evaluation using Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), with comparative analysis against traditional multi-regression models.

Main Results:

• The DNN model demonstrated accuracy and resilience in predicting typhoon damage.
• Validation metrics (MAE, RMSE) confirmed the model's performance.
• Comparative analysis indicated the proposed framework's superiority over traditional methods.

Conclusions:

• The proposed deep learning approach offers a novel and effective method for typhoon damage prediction.
• This framework can significantly aid disaster management, risk assessment, and insurance planning.
• Implementing this model can help reduce typhoon damage and bolster resilience to climate change impacts.