Empowering smart cities: Leveraging advanced forecasting models for proactive rainfall prediction and resilient urban planning

This study highlights the pivotal role of rainfall prediction within the dynamic landscape of smart cities. Accurate rainfall forecasts in such urban environments are foundational for bolstering infrastructure resilience, optimizing resource allocation, and ensuring the well-being of citizens. Employing an array of machine learning and statistical models—including Autoregressive Integrated Moving Average (ARIMA), Recurrent Neural Network (RNN), AdaBoost, Extreme Gradient Boosting (XGBoost), and Seasonal Autoregressive Integrated Moving Average (SARIMA)—this research delves into the prediction of rainfall patterns. Utilizing a comprehensive dataset spanning 116 years (1901–2016) sourced from the Pakistan Meteorological Department, rigorous preprocessing techniques addressed missing values and seasonal variations. Through meticulous segmentation into training and testing sets, the dataset facilitated robust model evaluation, employing diverse performance metrics such as Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Mean Squared Error (MSE), and Relative Root Mean Square Error (rRMSE). The analysis notably spotlights the outstanding performance of XGBoost and RNN among the models assessed. Specifically, XGBoost showcased exceptional metrics, with an RMSE of 0.1152 mm, MAE of 0.0834 mm, MAPE of 165.9995%, MSE of 0.0133 mm2, and rRMSE of 0.1573%, while RNN closely followed, with an RMSE of 0.1214 mm, MAE of 0.0893 mm, MAPE of 179.8016%, MSE of 0.0147 mm2, and rRMSE of 0.1657%. The integration of these advanced forecasting models into the framework of smart cities empowers urban planners and decision-makers to proactively address challenges posed by extreme weather events. By leveraging cutting-edge predictive techniques, smart cities can enhance their adaptability and responsiveness, ensuring sustainable development and improved quality of life for their inhabitants.