Air pollution, especially ground-level ozone, poses severe threats to human health and ecosystems. Accurate forecasting of ozone concentrations is essential for reducing its adverse effects. This study aims to use the functional time series approach to model ozone concentrations, a method less explored in the literature, and compare it with traditional time series and machine learning models. To this end, the ozone concentration hourly time series is first filtered for yearly seasonality using smoothing splines that lead us to the stochastic (residual) component. The stochastic component is modeled and forecast using a functional autoregressive model (FAR), where each daily ozone concentration profile is considered a single functional datum. For comparison purposes, different traditional and machine learning techniques, such as autoregressive integrated moving average (ARIMA), vector autoregressive (VAR), neural network autoregressive (NNAR), random forest (RF), and support vector machine (SVM), are also used to model and forecast the stochastic component. Once the forecast from the yearly seasonality component and stochastic component are obtained, both are added to obtain the final forecast. For empirical investigation, data consisting of hourly ozone measurements from Los Angeles from 2013 to 2017 are used, and one-day-ahead out-of-sample forecasts are obtained for a complete year. Based on the evaluation metrics, such as , root mean squared error (RMSE), and mean absolute error (MAE), the forecasting results indicate that the FAR outperforms the competitors in most scenarios, with the SVM model performing the least favorably across all cases.
Accurate short-term electricity demand forecasting (STDF) is critical not only for power system operational dependability but also for guiding strategic energy economic decision-making and promoting resilient, sustainable energy transitions. STDF informs market pricing, capacity planning, and regulatory policy by reducing forecasting uncertainty and enabling optimal resource allocation. This paper presents a novel hybrid forecasting framework that utilizes functional time series (FTS) modeling combined with advanced statistical anomaly detection methods to enhance forecasting accuracy in volatile power markets. We assess the performance of a variety of forecasting methods, including functional autoregressive (FAR), vector autoregressive (VAR), autoregressive moving average (ARMA), and machine learning models like autoregressive neural networks (ARNN) and k-nearest neighbors (KNN), in capturing complex temporal dependencies and structural dynamics of electricity demand. A key methodological novelty is the use of anomaly-aware preprocessing, which detects and corrects exogenous demand shocks that frequently distort time series models, particularly in deregulated or volatile energy markets. The FTS technique provides a concise yet expressive depiction of intra-day demand curves, allowing for both smooth functional variation and abrupt regime transitions. Empirical validation using high-frequency real-world data demonstrates that the proposed hybrid FTS framework consistently outperforms standard statistical and machine learning techniques in terms of prediction accuracy and resilience. This study makes a significant contribution to the literature on economic forecasting under uncertainty by combining advanced time series modeling with the demands of modern electricity markets. It emphasizes the need to integrate statistical rigor with machine learning flexibility to enhance predictive reliability, facilitate data-driven energy economics, and inform critical policy and investment decisions for long-term transitions.
Reliable short-term traffic flow prediction is crucial for intelligent transportation systems to enable real-time control, mitigate congestion, and improve urban mobility. However, traffic dynamics are inherently uncertain, temporally dependent, and subject to pronounced intraday variability, making accurate forecasting challenging. To address these issues, this study introduces a Functional AutoRegressive (FAR) model that represents daily traffic profiles as continuous stochastic functions rather than discrete observations, thereby preserving temporal continuity and capturing underlying symmetric structures. The model is developed using high-frequency traffic data collected at 15-min intervals from the Dublin Airport Link Road, Ireland, covering January 2022 to December 2024; data from 2022–2023 are used for model estimation, while 2024 data are reserved for one-day-ahead out-of-sample evaluation. A moving-window filtering technique is incorporated to enhance robustness by probabilistically identifying outliers and reducing noise. The proposed FAR approach is benchmarked against conventional models, including autoregressive (AR), autoregressive moving average (ARMA), nonparametric autoregressive (NPAR), and vector autoregressive (VAR) models. Empirical results demonstrate that the FAR model consistently achieves superior forecasting performance across all traffic conditions, yielding a full-day MAPE of 9.160% compared to 11.623% for the VAR model, along with lower MAE (76.772) and RMSE (131.767). It also performs best on both workdays and weekends, with MAPEs of 8.129% and 10.438%, respectively. Moreover, the model remains robust across peak and off-peak periods, effectively capturing both symmetric and asymmetric traffic variations while offering a more interpretable representation of intraday patterns. These findings suggest that functional time series modeling provides an effective and computationally efficient framework for traffic forecasting, with strong potential for application in next-generation intelligent transportation systems.