Smoothing Techniques for Improving COVID-19 Time Series Forecasting Across Countries
Accurate forecasting of COVID-19 case numbers is critical for timely and effective public health interventions. However, epidemiological data’s irregular and noisy nature often undermines the predictive performance. This study examines the influence of four smoothing techniques—the rolling mean, the...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-06-01
|
| Series: | Computation |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-3197/13/6/136 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849432000357728256 |
|---|---|
| author | Uliana Zbezhkhovska Dmytro Chumachenko |
| author_facet | Uliana Zbezhkhovska Dmytro Chumachenko |
| author_sort | Uliana Zbezhkhovska |
| collection | DOAJ |
| description | Accurate forecasting of COVID-19 case numbers is critical for timely and effective public health interventions. However, epidemiological data’s irregular and noisy nature often undermines the predictive performance. This study examines the influence of four smoothing techniques—the rolling mean, the exponentially weighted moving average, a Kalman filter, and seasonal–trend decomposition using Loess (STL)—on the forecasting accuracy of four models: LSTM, the Temporal Fusion Transformer (TFT), XGBoost, and LightGBM. Weekly case data from Ukraine, Bulgaria, Slovenia, and Greece were used to assess the models’ performance over short- (3-month) and medium-term (6-month) horizons. The results demonstrate that smoothing enhanced the models’ stability, particularly for neural architectures, and the model selection emerged as the primary driver of predictive accuracy. The LSTM and TFT models, when paired with STL or the rolling mean, outperformed the others in their short-term forecasts, while XGBoost exhibited greater robustness over longer horizons in selected countries. An ANOVA confirmed the statistically significant influence of the model type on the MAPE (<i>p</i> = 0.008), whereas the smoothing method alone showed no significant effect. These findings offer practical guidance for designing context-specific forecasting pipelines adapted to epidemic dynamics and variations in data quality. |
| format | Article |
| id | doaj-art-b02ce786d22e47dbace00545e2ca6d37 |
| institution | Kabale University |
| issn | 2079-3197 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Computation |
| spelling | doaj-art-b02ce786d22e47dbace00545e2ca6d372025-08-20T03:27:28ZengMDPI AGComputation2079-31972025-06-0113613610.3390/computation13060136Smoothing Techniques for Improving COVID-19 Time Series Forecasting Across CountriesUliana Zbezhkhovska0Dmytro Chumachenko1Scientific and Methodical Department for Quality Assurance of Educational Activities and Higher Education, Ivan Kozhedub Kharkiv National Air Force University, 61023 Kharkiv, UkraineMathematical Modelling and Artificial Intelligence Department, National Aerospace University “Kharkiv Aviation Institute”, 61072 Kharkiv, UkraineAccurate forecasting of COVID-19 case numbers is critical for timely and effective public health interventions. However, epidemiological data’s irregular and noisy nature often undermines the predictive performance. This study examines the influence of four smoothing techniques—the rolling mean, the exponentially weighted moving average, a Kalman filter, and seasonal–trend decomposition using Loess (STL)—on the forecasting accuracy of four models: LSTM, the Temporal Fusion Transformer (TFT), XGBoost, and LightGBM. Weekly case data from Ukraine, Bulgaria, Slovenia, and Greece were used to assess the models’ performance over short- (3-month) and medium-term (6-month) horizons. The results demonstrate that smoothing enhanced the models’ stability, particularly for neural architectures, and the model selection emerged as the primary driver of predictive accuracy. The LSTM and TFT models, when paired with STL or the rolling mean, outperformed the others in their short-term forecasts, while XGBoost exhibited greater robustness over longer horizons in selected countries. An ANOVA confirmed the statistically significant influence of the model type on the MAPE (<i>p</i> = 0.008), whereas the smoothing method alone showed no significant effect. These findings offer practical guidance for designing context-specific forecasting pipelines adapted to epidemic dynamics and variations in data quality.https://www.mdpi.com/2079-3197/13/6/136COVID-19 forecastingtime series smoothingSTL decompositionLightGBMLSTMTemporal Fusion Transformer |
| spellingShingle | Uliana Zbezhkhovska Dmytro Chumachenko Smoothing Techniques for Improving COVID-19 Time Series Forecasting Across Countries Computation COVID-19 forecasting time series smoothing STL decomposition LightGBM LSTM Temporal Fusion Transformer |
| title | Smoothing Techniques for Improving COVID-19 Time Series Forecasting Across Countries |
| title_full | Smoothing Techniques for Improving COVID-19 Time Series Forecasting Across Countries |
| title_fullStr | Smoothing Techniques for Improving COVID-19 Time Series Forecasting Across Countries |
| title_full_unstemmed | Smoothing Techniques for Improving COVID-19 Time Series Forecasting Across Countries |
| title_short | Smoothing Techniques for Improving COVID-19 Time Series Forecasting Across Countries |
| title_sort | smoothing techniques for improving covid 19 time series forecasting across countries |
| topic | COVID-19 forecasting time series smoothing STL decomposition LightGBM LSTM Temporal Fusion Transformer |
| url | https://www.mdpi.com/2079-3197/13/6/136 |
| work_keys_str_mv | AT ulianazbezhkhovska smoothingtechniquesforimprovingcovid19timeseriesforecastingacrosscountries AT dmytrochumachenko smoothingtechniquesforimprovingcovid19timeseriesforecastingacrosscountries |