New study uses weather to predict dengue outbreaks in Bangladesh
The system was 84.02% accurate, much better than earlier methods. A new way to pick the most important weather factors made the predictions even more precise, improving accuracy by 12.63% and reducing errors by 70.82%
Dengue is a disease spread by mosquitoes that causes serious health problems and affects the economy in tropical and subtropical areas, especially in developing countries like Bangladesh.
Better ways to predict and prepare for dengue outbreaks are urgently needed to improve mosquito control and make the best use of resources, says Nature.
The study, titled "Forecasting dengue in Bangladesh using meteorological variables with a novel feature selection approach" by Mahadee Al Mobin created a system to predict dengue cases using weather data, advanced machine learning, and special data processing techniques.
Keep updated, follow The Business Standard's Google news channel
The system was 84.02% accurate, much better than earlier methods. A new way to pick the most important weather factors made the predictions even more precise, improving accuracy by 12.63% and reducing errors by 70.82%. It also helped explain how much each weather factor affects dengue outbreaks.
The study found that relative humidity isn't useful for predicting dengue, but other weather factors are. For example, sunshine (hours of sunlight) has the biggest short-term effect on dengue cases, while rainfall is the most important over both short-term (8 months) and long-term (26–30 months) periods.
Using the best model, the study predicts a rise in dengue cases in 2024, starting in May and peaking at 24,000 cases per month in August. Cases will stay high through October and drop by December.
The study uses weather data and machine learning to predict dengue trends in Bangladesh. A special algorithm improves the accuracy and explains how different factors, like sunshine or rainfall, impact the disease.
The most accurate model used a decision tree and reached 84.02% accuracy. It showed that rainfall has both short- and long-term effects on dengue outbreaks, while sunshine has a more immediate impact. The findings predict a severe dengue outbreak in 2024, with cases peaking in late summer and early fall.
Bangladesh's current dengue prevention efforts are not well-coordinated, which makes it hard to control the disease and strains the healthcare system.
To fix this, the study suggests creating a national agency to use data and predictions to guide actions like mosquito control and healthcare planning. This approach could help the country manage dengue outbreaks and other health emergencies more effectively.
