Effects of Learning Rates and Optimization Algorithms on Forecasting Accuracy of Hourly Typhoon Rainfall: Experiments With Convolutional Neural Network

Uddin, Md. Jalal; Li, Yubin; Sattar, Md. Abdus; Nasrin, Zahan Most.; Lu, Chen

doi:10.1029/2021ea002168

Cited by 10 publications

(13 citation statements)

References 70 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A seasonal pattern means a repeated pattern in the time series, and the forecasting model can be affected by seasonality, as confirmed by the earlier study (Kwiatkowski et al, 1992). Recent studies (e.g., Prabhakaran, 2019;Uddin et al, 2022a) confirmed that when a time series had no seasonal effect, the outputs were more reliable. More information about the additive decomposition method is available in Seabold and Perktold (2010) (https://www.statsmodels.org/stable/index.html).…”

Section: Model Constructionsupporting

confidence: 52%

“…used the first 30 years (360 months) for training the model and the last 8 years (96 months) for evaluating the forecasting performance of the SVM, RF, CNN, and LSTM models. A grid search with the k-fold cross-validation method was employed in this study to find out the best parameters for the SVM, RF, CNN, and LSTM models, as proposed byUddin et al (2022aUddin et al ( , 2022b. The whole training dataset was cross-validated by using this method.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Climatic water balance forecasting with machine learning and deep learning models over Bangladesh

Uddin

Sattar

et al. 2022

Intl Journal of Climatology

Self Cite

View full text Add to dashboard Cite

Understanding the impact of input variables on black-box machine learning and deep learning models is necessary. Therefore, this study proposed SHapley Additive exPlanations (SHAP) values to address the problem of the interpretability of the output of the support vector machine (SVM), random forest (RF), convolutional neural network (CNN), and long short-term memory (LSTM) models to forecast climatic water balance (CWB) within 1-3 months lead-time.The current study uses two Koppen-Geiger climate zones over Bangladesh: the humid subtropical climate with dry winter and hot summer (Cwa) and the tropical climate (Af-Am). Monthly antecedent CWB, potential evapotranspira-

show abstract

Section: Model Constructionsupporting

confidence: 52%

mentioning

confidence: 99%

Climatic water balance forecasting with machine learning and deep learning models over Bangladesh

Uddin

Sattar

et al. 2022

Intl Journal of Climatology

Self Cite

View full text Add to dashboard Cite

show abstract

“…Future work could assess whether use of predictors that describe the spatial patterns of meteorological conditions, or use of other machine learning architectures such as convolutional neural networks (Uddin et al., 2022), can improve wind gust forecasting. It may also be useful to examine whether use of predictors from other reanalyzes or higher resolution datasets enhances forecast skill.…”

Section: Discussionmentioning

confidence: 99%

Short‐Term Forecasting of Wind Gusts at Airports Across CONUS Using Machine Learning

Coburn

Arnheim

Pryor

2022

Earth and Space Science

View full text Add to dashboard Cite

Short‐term forecasting of wind gusts, particularly those of higher intensity, is of great societal importance but is challenging due to the presence of multiple gust generation mechanisms. Wind gust observations from eight high‐passenger‐volume airports across the continental United States (CONUS) are summarized and used to develop predictive models of wind gust occurrence and magnitude. These short‐term (same hour) forecast models are built using multiple logistic and linear regression, as well as artificial neural networks (ANNs) of varying complexity. A suite of 19 upper‐air predictors drawn from the ERA5 reanalysis and an autoregressive (AR) term are used. Stepwise procedures instruct predictor selection, and resampling is used to quantify model stability. All models are developed separately for the warm (April–September) and cold (October–March) seasons. Results show that ANNs of 3–5 hidden layers (HLs) generally exhibit higher hit rates than logistic regression models and also improve skill with respect to wind gust magnitudes. However, deeper networks with more HLs increase false alarm rates in occurrence models and mean absolute error in magnitude models due to model overfitting. For model skill, inclusion of the AR term is critical while the majority of the remaining skill derives from wind speeds and lapse rates. A predictive ceiling is also clearly demonstrated, particularly for the strong and damaging gust magnitudes, which appears to be partially due to ERA5 predictor characteristics and the presence of mixed wind climates.

show abstract

“…In this study, we examine the formulations of SGD, Adam, RMSprop, AdaGrad, AdaDelta, and Nadam [17].…”

Section: The Optimizer Used In the Efficientnet Architecturementioning

confidence: 99%

Optimising Cataract Detection in Fundus Images through EfficientNet-Based Classification

Ibrahim,

Sabara,

Dirsam

et al. 2024

MEDINFTech

View full text Add to dashboard Cite

Turbidity of the lens of the eyeball that causes blindness or loss of vision is known as a cataract. By diagnosing the causes and symptoms of cataracts, early detection helps patients in prevention and treatment. The purpose of the research was to classify the image of the fundus into two classes: normal and cataract. The study also looked at how the optimizers for stochastic gradient descent, adaptive moment estimation, root mean square propagation, adaptive gradient algorithm, adaptive delta, and Nesterov-accelerated adaptive moment estimation stacked up against each other. We used the EfficientNet architecture in CNN and preprocessed the normal fundus and cataract fundus images by dividing each into training data (N = 80) and validation data (N = 20) from the Kaggle repository. We added test data from the normal fondus image (N =20) to see the accuracy of the results. We get 100% accuracy of training data, 87% and 77% validation data, and 100% and 95% test data.

show abstract

Effects of Learning Rates and Optimization Algorithms on Forecasting Accuracy of Hourly Typhoon Rainfall: Experiments With Convolutional Neural Network

Cited by 10 publications

References 70 publications

Climatic water balance forecasting with machine learning and deep learning models over Bangladesh

Climatic water balance forecasting with machine learning and deep learning models over Bangladesh

Short‐Term Forecasting of Wind Gusts at Airports Across CONUS Using Machine Learning

Optimising Cataract Detection in Fundus Images through EfficientNet-Based Classification

Contact Info

Product

Resources

About