Prediction of Infectious Disease Outbreaks (Cholera/Measles) in Iraq using SARIMA Models Optimized by the Reptile Search Algorithm (RSA)

Abstract

        Cholera and measles outbreaks were serious public health problems in Iraq, for which precise prediction models were necessary to support health planning and decision-making. This work proposed and evaluated a SARIMA model based on Reptile Search Algorithm (RSA) for forecasting cholera and measles data in Iraq. Monthly data from 2015 to 2024 were retrieved from WHO-EMRO and COSIT. The classical SARIMA model combined with the RSA algorithm was used to optimize the parameters. The improved model performed significantly better in predicting both cholera and measles on RMSE, MAE and MAPE levels. The RSA method showed time saving of 52.4% for cholera and 53.2% for measles against the traditional procedure. Tests for statistical significance (Diebold-Mariano test) revealed significant gains on the enhanced model with p < 0.001. Predictions of seasonal peaks for the 2025-26 period forecasted July-August for cholera and January-March for measles. These predictions were useful for proactive public health planning and allocation of resources. Evolutionary optimization algorithms could be synergistically coupled with conventional time series methods to improve the accuracy of the outbreak forecast for infectious diseases in resource-limited settings.