Mobil Ağ Baz İstasyonlarının Değişim Noktalarının Uzun Dönem Sezonsallık Tespiti için Hibrid Bir Algoritma

Abstract

Mobil ağlar için otomatik kapasite planlaması, geçmiş kalıpları kullanarak trafik talebinin uzun vadeli tahminini gerektirir. Doğru yatırım zamanına, doğru kapasite genişletme boyutuna karar vermede veya dışsal etkilere sahip tahmin algoritmalarının doğruluğunu iyileştirmede hem mevsimsel ayrıştırma hem de mevsimsel dönem tanımlama işlemleri karar doğruluğunu artırır. Bu çalışmada bu işlemleri, altyapısında parçalı Loess ile Mevsimsel Trend Ayrışımı (Seasonality Trend Decomposition with Loess – STL) ayrıştırması ve Prophet Kütüphanesi’nin Laplace önsele sahip regresyon yaklaşımını kullanan ve canlı ağ örnekleri üzerinde daha yüksek doğrulukla gerçekleştiren hibrid bir algoritma tasarlanmıştır. Her iki yöntemi de zayıf ve güçlü parçalarının farkındalığıyla birleştirmek ve değişim noktalarının benzerlik analizi ile tespit edilmesi üzerine geliştirilen çözüm, bu yöntemlerin tek başlarına elde ettiği ortalama başarımı yaklaşık %18,6 oranında artırmaktadır. Ayrıca çalışma kapsamında, problemin karmaşıklığını artıran ve ayrıştırma doğruluğunu azaltan bazı özel durumlar da sunulmuştur.

Keywords

• Mobil Ağ Sezonsallık Tespiti
• Değişim Noktası Farkındalığı
• Mobil Ağ Kapasite Planlama
• Parçalı Regresyon Analizi
• Zaman Serisi Bileşenlerine Ayırma

A Hybrid Algorithm for Changepoint Aware Long-Term Seasonality Detection of Mobile Network Base Stations

Abstract

Automated capacity planning for mobile networks requires long-term forecasting of traffic demand by using historical patterns. To decide the correct time of investment and correct capacity expansion size or to improve the accuracy of forecasting algorithms with exogenous features, both seasonal decomposition, and seasonal period identification improves decision accuracy. We design a hybrid algorithm to calculate these features on live network data with improved accuracy which uses piecewise Seasonality Trend Decomposition with Loess (STL) decomposition and Prophet library’s regression with Laplace prior under the hood. Combining both methods with the awareness of their weak and strong parts and leveraging overall output with changepoint and similarity analysis help us to improve our accuracy around 18.6% comparing the average of single usage of these methods. We also provide and present some special cases that increase problem complexity and decrease decomposition accuracy.

Keywords

• Mobile Network Seasonality Detection
• Changepoint Awareness
• Network Capacity Planning
• Piece-wise Regression
• Time Series Decomposition

Thanks

A part of this work has been conducted under the frame of the Celtic-Next AI4Green project where efficient and risk-aware energy saving algorithms are studied in collaboration. Calculating high season start and end dates next to the multiplicative impact of seasons is significant for risk minimization of RAN energy saving algorithms since contemporary solutions rely on short-term predictions over limited history.

References

1. Aminikhanghahi, S. & Cook, D. J. (2017). A Survey of Methods for Time Series Change Point Detection. Knowledge and Information Systems, 51(2), 339–367.
2. Balke, N. S. (1993). Detecting Level Shifts in Time Series. Journal of Business & Economic Statistics, 11, 81–92
3. Basseville, M. & Nikiforov, I. (1993). Detection of Abrupt Change Theory and Application, Prentice-Hall, ISBN: 0-13-126780-9.
4. Burg, G. J. J. & Williams, C. K. I. (2020). An Evaluation of Change Point Detection Algorithms.
5. Chen, H. & Zhang, N. (2015). Graph-based change-point detection. Annals of Statistics, 43(1), 139–176.
6. Cleveland, R. B., Cleveland, W. S., McRae, J. E., & Terpenning, I. J. (1990). STL: A seasonal-trend decomposition procedure based on loess. Journal of Official Statistics, 6(1), 3–33.
7. Cortez, P., Rio, M., Rocha, M. & Sousa, P. (2006). Internet traffic forecasting using neural networks. Proceedings of IEEE International Conference on Neural Networks, 2635–2642.
8. Dagum, E. B., & Bianconcini, S. (2016). Seasonal adjustment methods and real time trend-cycle estimation in Statistics for Social and Behavioral Sciences. Springer.

9. Downey, A. B. (2008). A novel changepoint detection algorithm.
10. Erp S.V., Oberski D.L., Mulder J. (2019). Shrinkage priors for Bayesian penalized regression. Journal of Mathematical Psychology, 89, 31–50.
11. Gould, P. G, Koehler, A. B., Ord, J.K., Snyder, R. D., Hyndman R.J. & Vahid-Araghi, F. (2008). Forecasting time series with multiple seasonal patterns, European Journal of Operational Research, 191, 207–222.
12. GSMA Intelligence. (2020). Mobile Economy Research Report. Available: https://www.gsma.com/mobileeconomy/wp-content/uploads/2020/03/GSMA_MobileEconomy2020_Global.pdf, Accessed on: Mar. 21, 2021.
13. Harvey, A. C. & Shephard, N. (1993), Structural time series models in Handbook of Statistics, Elsevier.
14. Hyndman, R. J. & Khandakar, Y. (2008). Automatic Time Series Forecasting: The forecast Package for R. Journal of Statistical Software, 27(3), 22.
15. Jaccard, P. (1912). The Distribution of the Flora in the Alphine Zone. New Phytologist, 11(2), 37–50.
16. Killick, R. & Eckley, I. A. (2014). changepoint: An R Package for Changepoint Analysis. Journal of Statistical Software, 58(3).
17. Lakshmanan, A. & Das, S. (2017). Two-stage models for forecasting time series with multiple seasonality.
18. Livera, A. M., Hyndman, R. J. & Snyder, R. D. (2011). Forecasting time series with complex seasonal patterns using exponential smoothing. Journal of the American Statistical Association, 106(496), 1513–1527
19. Nikravesh, A. Y., Ajila, S. A., Lung, C.-H. & Ding, W. (2016). An Experimental Investigation of Mobile Network Traffic Prediction Accuracy. Services Transactions on Big Data, 3(1), 1–16.
20. Page, E. S. (1954). Continuous inspection schemes. Biometrika 41.1/2, 100–115.
21. Rosner, B. (1975). On the detection of many outliers. Technometrics, 17(2), 221–227.
22. Sciancalepore, V., Samdanis, K., Costa-Perez, X., Bega, D., Gramaglia, M. & Banchs, A. (2017). Mobile traffic forecasting for maximizing 5G network slicing resource utilization, Proceedings of IEEE International Conference on Computer Communications.
23. Scott, A.J. & Knott, M. (1974). A Cluster Analysis Method for Grouping Means in the Analysis of Variance. Biometrics, 30(3), 507–512.
24. Seabold, Skipper & Perktold, J. (2010). statsmodels: Econometric and statistical modeling with python. Proceedings of the 9th Python in Science Conference.
25. Stan Development Team. (2020). Stan Modeling Language Users Guide and Reference Manual, 2.19.1. https://mc-stan.org
26. Taylor, S. J. & Letham, B. (2018). Forecasting at scale. The American Statistician, 72(1), 37–45.
27. Theodosiou M. (2011). Disaggregation & aggregation of time series components: A hybrid forecasting approach using generalized regression neural networks and the theta method. Neurocomputing, 74(6), 896–905.
28. Tikunov, D. & Nishimura, T. (2007). Traffic prediction for mobile network using Holt-Winter's exponential smoothing, Proceedingds of International Conference on Software, Telecommunications and Computer Networks, 310–314.
29. Yu, Y., Wang, J., Song, M. & Song, J. (2010). Network traffic prediction and result analysis based on seasonal ARIMA and correlation coefficient. Proceedings of 2010 International Conference on Intelligent System Design and Engineering Application, 1(1), 980–983.