Minimum Temperature Forecasting Using Gated Recurrent Unit

Subair, Hilma and Selvi, R. Pangayar and Vasanthi, R. and Kokilavani, S. and Karthick, V. (2023) Minimum Temperature Forecasting Using Gated Recurrent Unit. International Journal of Environment and Climate Change, 13 (9). pp. 2681-2688. ISSN 2581-8627

[thumbnail of Selvi1392023IJECC102397.pdf] Text
Selvi1392023IJECC102397.pdf - Published Version

Download (549kB)

Abstract

Aim: To forecast the monthly average Minimum Temperature (ºC) in Coimbatore district.

Study Design: Gated Recurrent Unit (GRU) has been employed to forecast the Minimum Temperature.

Place and Duration of Study: Time series data for average month wise Minimum Temperature from January 1982 to September 2022 was collected from Agro Climate Research Centre, TNAU for Coimbatore District.

Methodology: GRU which belongs to the field of deep learning has been employed to anticipate the average monthly Minimum Temperature by analyzing time series data from January 1982 to September 2022 in the district of Coimbatore. The model was trained using data from 1982 January to 2019 December and tested on data from 2020 January to 2022 September. After training and testing the algorithm was deployed to forecast Minimum Temperature for the lead time ahead.

Results: The GRU model generated RMSE and MAE scores of 0.694ºC and 0.523ºC, respectively, for Minimum Temperature. GRU model had a Willmott’s Index of Agreement (WI) value as 0.943 that is very close to 1. This demonstrates the effectiveness of the model built to effectively predict the Minimum Temperature. The study's evaluation of the RMSE, MAE, and Willmott Index value made it readily evident that the GRU model performed quite accurately for forecasting Minimum Temperature. Gated Recurrent Unit algorithm was used to forecast the Minimum Temperature from October 2022 till December 2023 that is for the next 15 months.

Item Type: Article
Subjects: Open Research Librarians > Geological Science
Depositing User: Unnamed user with email support@open.researchlibrarians.com
Date Deposited: 22 Sep 2023 13:12
Last Modified: 22 Sep 2023 13:12
URI: http://stm.e4journal.com/id/eprint/1544

Actions (login required)

View Item
View Item