<![CDATA[Acute Respiratory Infection (ARI) remains a significant public health problem in Indonesia, particularly in tropical coastal urban areas such as Semarang City. This study aims to analyze the effect of air quality (PM2.5) together with weather variables such as wind speed, temperature, humidity, and rainfall on ARI prevalence. Daily ARI case data based on ICD-10 codes from hospitals and meteorological data from BMKG for the period 2017–2024 were analyzed using Negative Binomial Regression for statistical inference, and artificial intelligence algorithms, namely Random Forest and SVM, for predictive modeling. The results show that ARI dynamics are predominantly influenced by temporal factors from previous cases, with weekly case variables (cases_ma_7 and total_kasus_lag7) emerging as the most important predictors. Environmental variables act as secondary modulators: temperature (TAVG_lag9) has an adverse effect, while wind speed (FF_AVG_lag0) has a direct positive impact. The Random Forest model demonstrates the best predictive performance with R² = 0.4443, RMSE = 4.98, MAE = 3.88, and MAPE = 35.77%. This study concludes that an ensemble learning approach using Random Forest is more accurate than SVM and statistical models for ARI prediction. Future research can explore hybrid artificial intelligence prediction and classification models that may serve as a basis for developing weather‑based early warning systems to support public health mitigation policies, particularly for ARI.]]>
Copyrights © 2026
