Authors: Koyu Mizutani, Haruki Mitarai, Kakeru Miyazaki, Soichiro Kumano, Toshihiko Yamasaki

Published on: February 03, 2024

Impact Score: 8.2

Arxiv code: Arxiv:2402.0215

Summary

• What is new: Development of linear regression models to predict seismic intensity distributions without requiring geographical information, capable of forecasting abnormal seismic intensity distributions.
• Why this is important: The need for accurate forecasting of earthquake damage and assessment of potential risks to save lives.
• What the research proposes: Using data-driven linear regression models based on earthquake parameters (location, depth, magnitude) to predict seismic intensity distributions.
• Results: The proposed model outperforms Ground Motion Prediction Equations (GMPEs) in correlation coefficient, F1 score, and MCC, demonstrating efficacy in predicting abnormal seismic intensity distributions.

Technical Details

Technological frameworks used: Linear regression

Models used: Regression and classification models, combined into a hybrid model

Data used: Seismic intensity data from 1,857 earthquakes in Japan (1997-2020)

Potential Impact

Emergency response and disaster management industries, insurance companies, construction and infrastructure sector

Want to implement this idea in a business?

We have generated a startup concept here: SeismoPredict.