relation: https://khub.utp.edu.my/scholars/19131/
title: Fires Hotspot Forecasting in Indonesia Using Long Short-Term Memory Algorithm and MODIS Datasets
creator: Kadir, E.A.
creator: Kung, H.T.
creator: Nasution, A.H.
creator: Daud, H.
creator: AlMansour, A.A.
creator: Othman, M.
creator: Rosa, L.
description: Vegetation fires are most common in South and Southeast Asian countries, including Indonesia. In addition to anthropogenic causes, climate change in the form of droughts is the biggest driver of fires in Indonesia. In particular, the peatlands in Indonesia are highly vulnerable to droughts with recurrent fires. In this study, we used a long short-term memory (LSTM) algorithm to predict the fire hotspots based on the 2010 to 2021 fire data. More than 700,000 fire hotspots from 2010 to 2021 have been collected and used as a training dataset to forecast fires for the year 2022. The LSTM algorithm successfully predicted 2022 fires with the minimum root mean squared error and high accuracy. Furthermore, the results of the 2022 prediction year matched the previous year�s fire data seasonally, with increasing fires from August to November. The study highlights the potential use of the LSTM algorithm for forecasting fires in Indonesia. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.
publisher: Springer International Publishing
date: 2023
type: Book
type: PeerReviewed
identifier:   Kadir, E.A. and Kung, H.T. and Nasution, A.H. and Daud, H. and AlMansour, A.A. and Othman, M. and Rosa, L.  (2023) Fires Hotspot Forecasting in Indonesia Using Long Short-Term Memory Algorithm and MODIS Datasets.     Springer International Publishing, pp. 589-602.  ISBN 9783031299162; 9783031299155     
relation: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85171518501&doi=10.1007%2f978-3-031-29916-2_35&partnerID=40&md5=23de4c2c760340a7871fa0bd35bb776d
relation: 10.1007/978-3-031-29916-2₃₅
identifier: 10.1007/978-3-031-29916-2₃₅