Hurricane-Induced Line Failure Risk Classification Using Machine Learning

Document Type

Conference Proceeding

Publication Date

12-8-2025

Department

Department of Mechanical and Aerospace Engineering

Abstract

This study examines the use of machine learning (ML) approaches for the predictive modeling of failure probability (FP) classes related to Atlantic hurricanes. FP was estimated from a collection of meteorological parameters and classified into Low, Medium, and High classes. The Synthetic Minority Oversampling Technique (SMOTE) was used to address the disparity in class. Both threshold-independent and threshold-based metrics, such as Accuracy, Precision, Recall, F1 Score, Average Precision Score (APS), Area Under the Receiver Operating Characteristic Curve (AUC), Cohen’s Kappa (Kappa), and Matthews Correlation Coefficient (MCC), were used to assess nine ML classifiers. On the test set, a number of models, including Decision Tree (DT), Gradient Boosting (GB), Random Forest (RF), and Logistic Regression (LR), performed flawlessly in classification. Top-performing models were subjected to 5-fold cross-validation and training time analysis to guarantee robustness. The study emphasizes the value of balancing procedures in multi-class hurricane failure classification as well as the efficacy of ensemble methods.

Publication Title

2025 IEEE 19th International Conference on Application of Information and Communication Technologies (AICT)

Share

COinS