Forecasting Malaria in Indian States: A Time Series Approach with R Shiny Integration.

Area of Science:

• Epidemiology
• Biostatistics
• Public Health

Background:

• Malaria poses a significant global health burden, necessitating advanced predictive modeling for effective control and prevention strategies.
• Accurate forecasting of malaria incidence is vital for resource allocation and intervention planning in endemic regions.

Purpose of the Study:

• To develop and evaluate time series models for forecasting malaria cases in eight key Indian states.
• To identify the most accurate and robust model for predicting malaria incidence dynamics.

Main Methods:

• Utilized various time series models: polynomial regression with seasonal components, log-transformed polynomial regression, lagged difference models, and ARIMA.
• Conducted comprehensive model fitting, residual analysis, and performance evaluation using RMSE and MAPE.
• Employed rolling forecast validation to assess predictive accuracy over time.

Main Results:

• The log-transformed polynomial regression model demonstrated superior accuracy and robustness compared to all other evaluated models across the eight states.
• Rolling forecast validation confirmed the consistent predictive superiority of the log-transformed model.
• An interactive R Shiny tool was developed for practical application of the predictive models.

Conclusions:

• Log-transformed polynomial regression is a highly effective method for malaria case forecasting in the studied Indian states.
• The developed R Shiny tool provides a valuable resource for public health officials and researchers for real-time malaria surveillance and decision-making.
• Appropriate statistical modeling significantly enhances malaria prediction capabilities, supporting improved public health interventions.