Issue |
E3S Web Conf.
Volume 585, 2024
5th International Conference on Environmental Design and Health (ICED2024)
|
|
---|---|---|
Article Number | 02006 | |
Number of page(s) | 9 | |
Section | Climate Change | |
DOI | https://doi.org/10.1051/e3sconf/202458502006 | |
Published online | 07 November 2024 |
Maximum temperature forecasting using deep learning algorithm by hyperparameter optimization
*1 Guru Ghasidas Vishwavidyalaya, Bilaspur (C.G), India
2 CSVTU, Bhilai(C.G), India
* Corresponding author: princy.matlani@gmail.com
The prediction of the daily temperature, an important meteorological variable, has been a topic of interest among researchers currently. The adverse impact of climate change on the livelihood of human beings makes it a contentious issue, hence the importance of accurate temperature predictions. In this paper, a global temperature change prediction model that adopts deep learning (DL) algorithms was presented which preprocess the Extreme-Weather Temperature Prediction Time Series Data by removing outliers using the standard deviation and normalizing the data. Statistical feature techniques are used for the extraction of characteristics, and forecasting is conducted using the Deep Belief Network (DBN) classifier. The proposed Egret Swarm Optimisation (ESO) method was used in training the multilayer perceptron (MLP) layer of the DBN. The success of the forecast is evaluated using mean absolute error (MAE), squared coefficient of correlation (R2), and root mean square error (RMSE). The results prove that the proposed model is better than as it has the lowest MAE (0.827), RMSE (0.892), the highest correlation (0.988), and the lowest Mean Absolute Relative Error (MARE) (0.126), showing a good linear relationship between the predicted and observed values, and low relative error (MARE). This makes it a significant advancement in temperature prediction.
© The Authors, published by EDP Sciences, 2024
This is an Open Access article distributed under the terms of the Creative Commons Attribution License 4.0, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.