Work

Analysis of the social impact of meteorological drought on Peninsular Malaysia based on the integration of precipitation satellites and sentiment analysis

Year: 2025
Type: article
Abstract: The integration of satellite data observations with other data sources and analytical tools, such as news media and machine-learning algorithms, will enhance the utilization of satellite information capabilities. This study validates 38 rain gauges with the merging satellite Tropical Rainfall Measuring Mission (TRMM) 3B43V7 (year 2000-2014) and its successor Global Precipitation Measuring Mission (GPM) 3IMERG (year 2015-2022) to monitor dry conditions in Peninsular Malaysia using the Standardized Precipitation Index (SPI). Seven English online news articles were used to collect 4086 label training headlines related to drought keywords. This study compares the performance of the text sentiment analysis model Naïve Bayes, Convolutional Neural Network (CNN), Bidirectional Encoder Representations from Transformers (BERT), and Long Short-Term Memory (LSTM). The evaluation focused on the accuracy and training time. The evaluation of TRMM-3B43V7 and GPM-3IMERG showed a significantly high correlation (R² = 0.873), and is a reliable source for providing a combined long-term precipitation record and drought study in Peninsular Malaysia. The results of sentiment analysis using the LSTM method showed a high accuracy of 0.804, whereas the duration of training was significantly lower than that of the other models. Finally, the trained LSTM model was used to predict the sentiment on a separate headline input dataset and compare the expected sentiment result with drought-affected areas from the satellite SPI-12 timescale outputs. During drought or no-drought events, water remains a significant problem in most economic growth contributor states, such as Selangor and Penang. Hence, sustainable water management must be implemented to improve socioeconomic development.
Language: English
Open Access status: hybrid
Estimated APC paid: $3,200
OpenAlex ID: W4407599105
FWCI: 2.102
Cites: 30
Cited by: 5
Related to: 10