Binary vs Multi-class with Gaussian Filter on Typhoon Image Classification for Intensity Prediction DOI Open Access

Syamala Jayasree,

K. R. Ananthapadmanaban

International Journal of Electronics and Communication Engineering, Journal Year: 2024, Volume and Issue: 11(12), P. 245 - 257

Published: Dec. 31, 2024

Strong meteorological events Tropical Cyclones (TCs) pose serious risks to coastal ecosystems and communities. Their strength is usually categorized using a variety of metrics, including wind speed, pressure, rainfall since it directly corresponds with the possibility damage fatalities. An accurate classification TC severity essential for disaster preparedness, response plans, mitigation initiatives. Support vector machines (SVM) {function category}, K-Nearest Neighbors (KNN) {lazy Bayesian networks {Bayes Random forests {Ensemble decision trees {Tree Category} are among machine learning classifiers whose performances compared in this study binary multi-class configurations by Gaussian image processing technique. Performance measures, time complexity, ROC, PRC, accuracy, precision, recall, F-measure, were examined. The results indicate that Multi-class SVM Forest consistently outperform other models across most achieving highest accuracy (0.88) superior ROC (0.97) PRC (0.94-0.95) scores. However, exhibited significantly higher particularly SVM.

Language: Английский

Cascaded neural network surrogate modeling for real-time decision-making in long-distance water supply distribution DOI Creative Commons
Lin Shi, Jian Zhang, Sheng Chen

et al.

Engineering Applications of Computational Fluid Mechanics, Journal Year: 2025, Volume and Issue: 19(1)

Published: Jan. 16, 2025

Effective water distribution in long-distance supply systems requires precise control over pump station operations and flow-regulating elements, such as speeds valve openings, typically achieved through hydraulic models. However, traditional models are time-intensive to develop require frequent calibration, limiting their practicality for real-time applications. This paper presents a cascaded neural network (CNN) model that integrates classification regression components serve an efficient surrogate decision-making. In the proposed CNN model, component identifies number of pumps needed meet system flow demands, while predicts target values openings. Considering nonlinear relationship between rate regulating error was introduced evaluation metric via Orthogonal-Triangular (QR) decomposition. The model's performance robustness were validated using data from actual system, including analyses its sensitivity uncertainties reservoir level measurements. Results demonstrate achieves more accurate predictions compared pure networks. Furthermore, uncertainty analysis reveals is less affected by measurement errors, it sensitive underscoring importance monitoring practical

Language: Английский

Citations

0
City-Scale High-Resolution Flood Nowcasting Based on High-Performance Hydrodynamic Modelling DOI
Boliang Dong,

Chao Tan,

Bensheng Huang

et al.

Published: Jan. 1, 2025

Language: Английский

Citations

0
Binary vs Multi-class with Gaussian Filter on Typhoon Image Classification for Intensity Prediction DOI Open Access

Syamala Jayasree,

K. R. Ananthapadmanaban

International Journal of Electronics and Communication Engineering, Journal Year: 2024, Volume and Issue: 11(12), P. 245 - 257

Published: Dec. 31, 2024

Strong meteorological events Tropical Cyclones (TCs) pose serious risks to coastal ecosystems and communities. Their strength is usually categorized using a variety of metrics, including wind speed, pressure, rainfall since it directly corresponds with the possibility damage fatalities. An accurate classification TC severity essential for disaster preparedness, response plans, mitigation initiatives. Support vector machines (SVM) {function category}, K-Nearest Neighbors (KNN) {lazy Bayesian networks {Bayes Random forests {Ensemble decision trees {Tree Category} are among machine learning classifiers whose performances compared in this study binary multi-class configurations by Gaussian image processing technique. Performance measures, time complexity, ROC, PRC, accuracy, precision, recall, F-measure, were examined. The results indicate that Multi-class SVM Forest consistently outperform other models across most achieving highest accuracy (0.88) superior ROC (0.97) PRC (0.94-0.95) scores. However, exhibited significantly higher particularly SVM.

Language: Английский

Citations

0