A MACHINE LEARNING APPROACH TO SEISMIC RISK ASSESSMENT OF REINFORCED CONCRETE STRUCTURES

Date

2023-8-29

Author

Beyhan, M. Tunahan

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

598
views

435
downloads

Cite This

This research aims to bring a novel approach to seismic risk assessment of low-rise reinforced concrete (RC) buildings by leveraging advanced machine learning techniques. Utilizing a meticulously compiled 4699 RC building dataset, evaluated according to the “Guidelines for the Assessment of Risky Buildings” (GARB) code, the study deploys a range of machine learning algorithms to predict the seismic risk status of these structures. The dataset comprises 13 independent variables and one dependent variable, the “Risk Status.” The data underwent extensive preprocessing, including cleaning, normalization, feature selection, engineering, and applying the Synthetic Minority Over-sampling Technique (SMOTE) to address the class unbalance. Nine machine learning algorithms - Decision Tree, Support Vector Machine, Random Forest, Logistic Regression, K-Nearest Neighbors, Naive Bayes, XGBoost, Artificial Neural Network, and LightGBM - were evaluated using essential metrics such as precision, recall, F1-score, and the area under the Receiver Operating Characteristic curve (AUC-ROC). Following hyperparameter tuning and cross-validation, the XGBoost and NaiveBayes models emerged as the most effective algorithms according to different scoring metrics. Furthermore, this work identifies the most influential building characteristics for seismic risk prediction. The research represents a significant contribution to seismic risk assessment methodologies, highlighting the potential of machine learning in this field.

Subject Keywords

Seismic Risk Assessment, Reinforced Concrete Buildings, Earthquake Engineering, Data Science, Machine Learning

URI

https://hdl.handle.net/11511/105315

Collections

Graduate School of Natural and Applied Sciences, Thesis