Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Predicting building damage grade by earthquake: a Bayesian Optimization-based comparative study of machine learning algorithms

Predicting building damage grade by earthquake: a Bayesian Optimization-based comparative study... This study compares Bayesian Optimization-based machine learning systems that anticipate earthquake-damaged buildings and to evaluates building damage classification models. Using metrics, this study evaluates Random Forest, ElasticNet, and Decision Tree algorithms. This study showed damage level asymmetry. Fifth grade is the most prevalent and first grade the least. The class imbalance makes estimating building damage grades difficult, emphasizing the necessity for careful modeling. Bayesian Optimization optimizes machine learning algorithm hyperparameters to solve this problem. The optimization technique maximizes the receiver operating characteristic curve (AUROC), which measures the models’ ability to discern between damage levels. Convergence shows that Bayesian Optimization improves model discrimination. The optimized models classified building damage grades with an AUROC of 0.9952. Comparing machine learning algorithms yields insights. The ElasticNet model predicts building damage grade with 92.56% test accuracy and 92.67% train accuracy. With 89.39% test accuracy and 99.82% train accuracy, the Random Forest model performs well. The Decision Tree model has 89.19% test and 99.94% train accuracy. Mean Squared Error (MSE) shows that the Random Forest model makes more accurate predictions than the others. In this study, machine learning techniques forecast building deterioration. Research should address class imbalance, because it affects model performance. Bayesian Optimization helps models acquire data patterns, improving classification accuracy. This study shows that machine learning and optimization can forecast building seismic damage grades. The proposed model can successfully use for earthquake risk assessment and mitigation. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Asian Journal of Civil Engineering Springer Journals

Predicting building damage grade by earthquake: a Bayesian Optimization-based comparative study of machine learning algorithms

Loading next page...
 
/lp/springer-journals/predicting-building-damage-grade-by-earthquake-a-bayesian-optimization-9yw0IM0uhN

References (35)

Publisher
Springer Journals
Copyright
Copyright © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2023. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
ISSN
1563-0854
eISSN
2522-011X
DOI
10.1007/s42107-023-00771-6
Publisher site
See Article on Publisher Site

Abstract

This study compares Bayesian Optimization-based machine learning systems that anticipate earthquake-damaged buildings and to evaluates building damage classification models. Using metrics, this study evaluates Random Forest, ElasticNet, and Decision Tree algorithms. This study showed damage level asymmetry. Fifth grade is the most prevalent and first grade the least. The class imbalance makes estimating building damage grades difficult, emphasizing the necessity for careful modeling. Bayesian Optimization optimizes machine learning algorithm hyperparameters to solve this problem. The optimization technique maximizes the receiver operating characteristic curve (AUROC), which measures the models’ ability to discern between damage levels. Convergence shows that Bayesian Optimization improves model discrimination. The optimized models classified building damage grades with an AUROC of 0.9952. Comparing machine learning algorithms yields insights. The ElasticNet model predicts building damage grade with 92.56% test accuracy and 92.67% train accuracy. With 89.39% test accuracy and 99.82% train accuracy, the Random Forest model performs well. The Decision Tree model has 89.19% test and 99.94% train accuracy. Mean Squared Error (MSE) shows that the Random Forest model makes more accurate predictions than the others. In this study, machine learning techniques forecast building deterioration. Research should address class imbalance, because it affects model performance. Bayesian Optimization helps models acquire data patterns, improving classification accuracy. This study shows that machine learning and optimization can forecast building seismic damage grades. The proposed model can successfully use for earthquake risk assessment and mitigation.

Journal

Asian Journal of Civil EngineeringSpringer Journals

Published: Jan 1, 2024

Keywords: Building damage grade prediction; Machine learning algorithms; Bayesian optimization; Class imbalance; Comparative analysis

There are no references for this article.