Pandemic-Driven digitalization in Türkiye: a hybrid regression, ARIMA–SARIMA and machine learning analysis of mobile banking
Abstract
This research scrutinizes the abrupt behavioral shifts surrounding digital financial platforms in the Turkish sector, utilizing a multidimensional forecasting apparatus that fuses classical regression, seasonal econometrics, and tree-based predictive learners. Fifty-three quarters of aggregate usage logs (late-2011 to late-2024) were processed via a tiered methodology: foundational polynomial estimators mapped the continuous growth baseline, while complex seasonal autoregressive models (SARIMA) absorbed cyclic temporal echoes. To address erratic nonlinear surges, ensemble protocols—specifically Random Forest and XGBoost—were subsequently deployed alongside an innovative composite framework that mitigates structured econometric residuals via iterative gradient boosting. Our quantitative results isolate a violent, statistically verified inflection point triggered during the second quarter of 2020, an anomaly that persistently overwhelmed standard linear projection boundaries. Comparatively, the integrated SARIMA–XGBoost construct achieved unmatched tracking accuracy (RMSE: 214,586; MAE: 122,847; MAPE: 2.6%; 𝑅2: 0.995), dramatically eclipsing the precision of isolated ARIMA frameworks. Ultimately, this analysis corroborates the assertion that melding standard time-series inference with algorithmic error-correction produces extraordinarily resilient predictive tools, particularly when modeling markets destabilized by sudden macroeconomic shocks.
Keywords
Mobile Banking, COVID-19, Ordinary Least Squares Regression, ARIMA, Hybrid Statistical–Machine Learning
Ethical Statement
References
- Akaike H., A new look at the statistical model identification, IEEE Transactions on Automatic Control, 19(6), 716–723, 1974.
- Alshammary M., Alshammary F., and Alqahtani M., Digital transformation of the financial sector during the COVID-19 pandemic, Journal of Innovation&Knowledge, 7(3), 100204, 2022.
- Box G. E. P., Jenkins G. M., Reinsel G. C., and Ljung G. M., Time Series Analysis: Forecasting and Control (5th ed.). Wiley, 2015.
- Breiman L., Random forests, Machine Learning, 45(1), 5–32, 2001.
- Bretscher O., Linear Algebra with Applications (5th ed.). Otto Bretsch, 2013.
- Chen T., and Guestrin C., XGBoost: A scalable tree boosting system, in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 785–794, 2016.
- Demirg¨uc¸-Kunt A., Pedraza A., and Ruiz-Ortega C., Banking sector performance during the COVID-19 crisis, Journal of Banking & Finance, 133, 106305, 2021.
- Dickey D. A., and Fuller W. A., Distribution of the estimators for autoregressive time series with a unit root, Journal of the American Statistical Association, 74(366a), 427–431, 1979.
- Dietz W., and Santos-Burgoa C., Obesity and its implications for COVID-19 mortality, Obesity, 28(6), 1005, 2020.
- Durbin J., and Watson G. S., Testing for serial correlation in least squares regression. II, in Breakthroughs in Statistics: Methodology and Distribution, 260–266. Springer, 1992.