Time-series decomposition and modeling of dengue cases in Malaysia, 2022-2024: a nationwide observational study.

backgroundThis study aimed to examine the temporal dynamics of dengue cases in Malaysia from 2022 to 2024 using seasonal-trend decomposition and time-series modeling.

methodsWeekly dengue case counts from the national registry were analyzed across all states using seasonal-trend decomposition using LOESS (STL) to separate trend, seasonal, and irregular components. Autoregressive integrated moving average (ARIMA) and seasonal ARIMA (SARIMA) models were fitted to validate temporal structures, with model selection based on the Akaike information criterion (AIC), corrected AIC, and Bayesian information criterion. Diagnostic checks, including residual analysis and Ljung-Box testing, were performed to ensure model adequacy.

resultsDengue incidence showed marked heterogeneity across states. STL decomposition indicated that long-term trends contributed more strongly to case dynamics than seasonality in most states, although seasonal influences were significant in the states of Kedah and Kelantan. Seasonal peak timing varied between states, highlighting differences in epidemic cycles. ARIMA and SARIMA modeling confirmed that no single temporal structure could adequately represent all states; while some series were well fitted by simple ARIMA models, others required seasonal adjustments. Residual diagnostics demonstrated that the selected models were statistically adequate.

conclusionDengue dynamics in Malaysia are shaped by both trend and seasonal components, with considerable variation across states. Combining STL decomposition with ARIMA/SARIMA modeling strengthens the evidence base for state-specific forecasting and proactive vector control. Tailoring surveillance systems and interventions to local temporal patterns.