Heat Wave Prediction Using Recurrent Neural Networks Based on Deep Learning

Sarita Byagar, Araddhana Arvind Deshmukh, Kirti Wanjale, Vinod S. Wadne, Nidhi Ranjan, Rupali Gangarde

International Journal of Intelligent Systems and Applications in Engineering · 03 September 2023

In one sentence. LSTM predicts local heatwave occurrence from historical temperature, humidity and wind data, outperforming traditional statistical models in accuracy and F1.

Abstract

There are significant threats to agriculture, the environment, and human health as a result of the increasing frequency and intensity of heatwaves. The success of mitigation and adaptation strategies depends on accurate heatwave forecast. In this paper, we suggest a deep learning-based method for local heatwave prediction using recurrent neural networks (RNNs). With the help of historical meteorological data, such as temperature, humidity, wind speed, and other pertinent variables, the suggested model investigates the intricate temporal patterns related to the occurrence of heatwaves. The RNN design uses the Long Short-Term Memory (LSTM) to retain long-term dependencies and quickly process sequential data. The training data are used to construct the RNN model, then grid search and cross-validation techniques are used to improve its hyperparameters. Several evaluation criteria are employed to assess the model's performance, including accuracy, precision, recall, and F1-score. The results show that the deep learning-based method for local heatwave prediction works well. In terms of accuracy, recall, precision, and F1-score, the model performs admirably. It also performs better than traditional statistical models and shows the efficacy of deep learning approaches in recognising the complex spatiotemporal patterns related to heatwave occurrences.

Labels

Labels and summary are generated by Claude and reviewed weekly. Wrong? Edit this entry.