Deep Learning to implement a Statistical Weather Forecast for the Andean City of Quito

Canar, Ricardo Llugsi; Fontaine, Allyx; Morillo, Pablo Lupera; Yacoubi, Samira El

doi:10.1109/andescon50619.2020.9272106

Cited by 3 publications

(3 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The LSTM consists of memory blocks that allow it to store long-term memory well. The LSTM memory block consists of a memory cell and three gate units called forget gate, input gate, and output gate (Canar et al, 2020;Mimboro et al, 2021) three unit gates will control the flow of data that will enter and exit the memory cell. The LSTM memory block architecture is shown in Figure 1.…”

Section: Bidirectional Long-short Term Memory (Bilstm)mentioning

confidence: 99%

Improving Accuracy of Daily Weather Forecast Model at Soekarno-Hatta Airport Using BILSTM with SMOTE and ADASYN

Danitasari,

Ryan,

Handoko

et al. 2024

jppipa, pendidikan ipa, fisika, biologi, kimia

View full text Add to dashboard Cite

Bidirectional LSTM (BiLSTM) is an extension of LSTM which can improve model efficiency and accuracy in classification scenarios based on time series data or longer time series data repeatedly. This research uses the BiLSTM algorithm to build a daily weather forecast model at Soekarno-Hatta Airport. The model built will assist forecasters in making weather forecasts on a local scale. This research is expected to be implemented and able to increase the verification value of Soekarno-Hatta Airport weather forecasts to support flight safety in Indonesia. The dataset used is hourly surface air weather parameter data (synoptic data) of Soekarno-Hatta Meteorological Station for the period January 2018 - December 2022. There is an imbalance in the data set, so the SMOTE and ADASYN techniques are used to handle the problem. The output of this research is weather conditions categorised into sunny, sunny cloudy, cloudy, light rain, moderate rain, heavy rain, and thunder rain. The results obtained will go through model verification and evaluation by finding the accuracy value by comparing the weather forecast model output with actual weather data using a multi-category contingency table. The BiLSTM - ADASYN model obtained the highest average accuracy value compared to other models, which was 83.2%.

show abstract

Section: Bidirectional Long-short Term Memory (Bilstm)mentioning

confidence: 99%

Improving Accuracy of Daily Weather Forecast Model at Soekarno-Hatta Airport Using BILSTM with SMOTE and ADASYN

Danitasari,

Ryan,

Handoko

et al. 2024

jppipa, pendidikan ipa, fisika, biologi, kimia

View full text Add to dashboard Cite

show abstract

“…the past years has gained a lot of attention even has challenged the performance of NWP models [1]. Deep Learning based models have been previously successfully applied for forecasting tasks in a range of application domains including crop yield [2], solar irradiance [3], traffic [4] and weather [5][6][7][8][9][10][11][12]. The weather datasets are time-series with observations of a weather element at each time step.…”

Section: Introductionmentioning

confidence: 99%

SAR-UNet: Small Attention Residual UNet for Explainable Nowcasting Tasks

Renault

Mehrkanoon

2023

2023 International Joint Conference on Neural Networks (IJCNN)

View full text Add to dashboard Cite

The accuracy and explainability of data-driven nowcasting models are of great importance in many socio-economic sectors reliant on weather-dependent decision making. This paper proposes a novel architecture called Small Attention Residual UNet (SAR-UNet) for precipitation and cloud cover nowcasting. Here, SmaAt-UNet is used as a core model and is further equipped with residual connections, parallel to the depthwise separable convolutions. The proposed SAR-UNet model is evaluated on two datasets, i.e., Dutch precipitation maps ranging from 2016 to 2019 and French cloud cover binary images from 2017 to 2018. The obtained results show that SAR-UNet outperforms other examined models in precipitation nowcasting from 30 to 180 minutes in the future as well as cloud cover nowcasting in the next 90 minutes. Furthermore, we provide additional insights on the nowcasts made by our proposed model using Grad-CAM, a visual explanation technique, which is employed on different levels of the encoder and decoder paths of the SAR-UNet model and produces heatmaps highlighting the critical regions in the input image as well as intermediate representations to the precipitation. The heatmaps generated by Grad-CAM reveal the interactions between the residual connections and the depthwise separable convolutions inside of the multiple depthwise separable blocks placed throughout the network architecture.

show abstract

“…Such neural networks have been successfully applied to several distinct domains such as financial market forecasting [Moghar andHamiche 2020] [de Caux et al 2020], weather forecasting [Tukymbekov et al 2021] [Cañar et al 2020, Covid-19 transmission [Patidar et al 2021], prediction of failures in aeronautical components [Chui et al 2021], among others. Some issues exist when relying on artificial neural networks since they usually require a training phase in addition to being sensitive to the high number of hyperparameters [Elsworth and Güttel 2020].…”

Section: Introductionmentioning

confidence: 99%

On the benefits of automated tuning of hyper-parameters: an experiment related to temperature prediction on UAV computers

Maximiano¹,

Júnior²,

Shiguemori³

2022

Anais Do XIX Encontro Nacional De Inteligência Artificial E Computacional (ENIAC 2022)

View full text Add to dashboard Cite

Finding the best configuration of a neural network to solve a problem has been challenging given the numerous possibilities of values of the hyper-parameters. Thus, tuning of hyper-parameters is one important approach and researchers suggest doing this automatically. However, it is important to verify when it is suitable to perform automated tuning which is usually very costly financially and also in terms of hardware infrastructure. In this study, we analyze the advantages of using a hyper-parameter optimization framework as a way of optimizing the automated search for hyper-parameters of a neural network. To achieve this goal, we used data from an experiment related to temperature prediction of computers embedded in unmanned aerial vehicles (UAVs), and the models Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) to perform these predictions. In addition, we compare the hyper-parameter optimization framework to the hyper-parameter exhaustive search technique varying the size of the training dataset. Results of our experiment shows that designing a model using a hyper-parameter optimizer can be up to 36.02% better than using exhaustive search, in addition to achieving satisfactory results with a reduced dataset.

show abstract

Deep Learning to implement a Statistical Weather Forecast for the Andean City of Quito

Cited by 3 publications

References 18 publications

Improving Accuracy of Daily Weather Forecast Model at Soekarno-Hatta Airport Using BILSTM with SMOTE and ADASYN

Improving Accuracy of Daily Weather Forecast Model at Soekarno-Hatta Airport Using BILSTM with SMOTE and ADASYN

SAR-UNet: Small Attention Residual UNet for Explainable Nowcasting Tasks

On the benefits of automated tuning of hyper-parameters: an experiment related to temperature prediction on UAV computers

Contact Info

Product

Resources

About