Acil servise başvuru sayısının zaman serisi analiz ve makine öğrenmesi yöntemleri ile tahmin edilmesine yönelik bir uygulama

Abstract

Günümüzde, acil sağlık servislerine yönelik talepler, salgın, deprem vb. doğal afetler ile patlamalar gibi durumlarda olağanüstü artış göstermektedir. Söz konusu talebin doğru bir şekilde tahmin edilmesi, acil servislere başvuracak kişi sayısının belirlenmesi ve ilgili kaynak planlamalarının etkin şekilde gerçekleştirilmesini sağlayacağından olağanüstü durumlar için kriz yönetim sürecinde kolaylık sağlayacaktır. Bu çalışmada, bir acil servise başvuru sayısının tahmini amaçlanmaktadır. Ele alınan, mevsimsel özelliklere sahip verilere yönelik olarak, zaman serisi analizi yöntemlerinden SARIMA, HoltWinters ve ayrıştırma; makine öğrenmesi yöntemlerinden rastgele ağaç ve rastgele orman teknikleri kullanılmıştır. Tahmin çalışması için Ankara’ da yer alan bir hastanenin 396 günlük “başvuran hasta sayısı” verisi kullanılmıştır. Her bir yöntemdeki tahminler, yedi, on beş ve otuz günlük olarak yapılmıştır. Talep tahmini yöntemlerinden en başarılı yöntemin belirlenebilmesi için korelasyon, düzeltilmiş RKARE ve ortalama mutlak yüzde hatası değerlerinden faydalanılmıştır. Yapılan analizlerde SARIMA yönteminin, acil servise yapılacak başvuru sayısının tahmin edilmesinde diğer yöntemlere göre daha etkili sonuçlar verdiği görülmüştür. Ayrıca, acil servislere yapılan başvuruların sürekli değişen, dinamik bir yapıya sahip olmasının bir sonucu olarak, tahmin edilen gün sayısındaki değişimin tahmin değerleri üzerinde önemli etkisi olduğu da anlaşılmıştır.

Keywords

• Talep tahmini
• Acil servis
• Zaman serisi analizi
• SARIMA
• Holt-Winters
• Makine öğrenmesi

An application for forecasting the number of applications to the emergency department with time series analysis and machine learning methods

Abstract

Today, the demands for emergency health services show an extraordinary increase in cases such as epidemics, earthquakes, natural disasters, and explosions. The accurate estimation of the demand in question will facilitate the crisis management process for extraordinary situations, as it will enable the determination of the number of people who will apply to the emergency services and the effective realization of the relevant resource planning. In this study, it is aimed to estimate the number of applications to an emergency department. For the seasonal data, SARIMA, Holt-Winters and decomposition, which are among the time series analysis methods; Random tree and random forest techniques from machine learning methods are used. For this forecasting study, 396-day "number of patients admitted" data of a hospital located in Ankara is used. Forecasts in each method are performed for seven, fifteen, and thirty days. Correlation corrected square root and average absolute percentage error values are used to determine the most successful one among demand forecasting methods. In the analyzes made, it is observed that SARIMA method gives more effective results than others in forecasting the number of applications to the emergency department. In addition, because of the constantly changing and dynamic nature of the applications made to the emergency services, it is understood that the change in the forecasted number of days has a significant effect on the resulting forecast values.

Keywords

• Demand forecast
• Emergency service
• Time series analysis
• SARIMA
• Holt-Winters
• Machine learning

References

1. [1] Afilal M, Yalaoui F, Dugardin F, Amodeo L, Laplanche D, Blua P. “Forecasting the emergency department patients flow”. Journal of Medical Systems, 40(7), 1-18, 2016.
2. [2] Harrou F, Dairi A, Kadri F, Sun Y. “Effective forecasting of key features in hospital emergency department: Hybrid deep learning-driven methods”. Machine Learning with Applications, 7, 1-13, 2022.
3. [3] Ramgopal S, Pelletier JH, Rakkar J, Horvat CM. “Forecast modeling to identify changes in pediatric emergency department utilization during the COVID-19 pandemic”. The American Journal of Emergency Medicine, 49, 142-147, 2021.
4. [4] Becerra M, Jerez A, Aballay B, Garcés HO, Fuentes A. “Forecasting emergency admissions due to respiratory diseases in high variability scenarios using time series: A case study in Chile”. Science of the Total Environment, 706, 1-11, 2020.
5. [5] Bekker R, Broek M, Koole G. “Modeling COVID-19 hospital admissions and occupancy in the Netherlands”. European Journal of Operational Research, 304(1), 207-218, 2023.
6. [6] Dam V, Zelis P, Kuijk S, Linkens A, Brüggeman R, Spaetgens B, Horst I, Stassen P. “Performance of prediction models for short-term outcome in COVID-19 patients in the emergency department: a retrospective study”. Annals of Medicine, 53(1), 402-409, 2021.
7. [7] Ozudogru AG, Gorener A. “Method selection for demand forecasting: Application in a private hospital”. International Journal of Decision Sciences & Applications, 1(1), 13-22, 2020.
8. [8] Jones SA, Joy MP, Pearson J. “Forecasting demand of emergency care”. Health Care Management Science 2002, 5(4), 297-305, 2002.

9. [9] Noel G, Bonte N, Persico N, Bar C, Luigi S, Roch, A, Viudesa G. “Real-time estimation of inpatient beds required in emergency departments”. European Journal of Emergency Medicine, 26(6), 440-445, 2019.
10. [10] McCoy TH, Pellegrini AM, Perlis RH. “Assessment of timeseries machine learning methods for forecasting hospital discharge volume”. JAMA Network Open, 1(7), 1-9, 2018.
11. [11] Jones SS, Thomas A, Evans RS, Welch SJ, Haug PJ, Snow GL. “Forecasting daily patient volumes in the emergency department”. Academic Emergency Medicine, 15(2), 159-170, 2008.
12. [12] Jones SS, Evans RS, Allen TL. Thomas A, Haug PJ, Welch SJ, Snow GL. “A multivariate time series approach to modeling and forecasting demand in the emergency department”. Journal of Biomedical Informatics, 42(1), 123-139, 2009.
13. [13] Luo W, Cao J, Gallagher M, Wiles J. “Estimating the intensity of ward admission and its effect on emergency department access block”. Statistics in medicine, 32(15), 2681-2694, 2013.
14. [14] Bergs J, Heerinckx P, Verelst S, “Knowing what to expect, forecasting monthly emergency department visits: a timeseries analysis”. International Emergency Nursing, 22(2), 112-115, 2014.
15. [15] Peck JS, Benneyan JC, Nightingale DJ, Gaehde SA. “Predicting emergency department inpatient admissions to improve same‐day patient flow”. Academic Emergency Medicine, 19(9), 1045-1054, 2012.
16. [16] Barrett TW, Martin AR, Storrow AB, Jenkins CA, Harrell Jr FE, Russ S, Darbar D. “A clinical prediction model to estimate risk for 30-day adverse events in emergency department patients with symptomatic atrial fibrillation”. Annals of Emergency Medicine, 57(1), 1-12, 2011.
17. [17] Sariyer G. “Acil servislerde talebin zaman serileri modelleri ile tahmin edilmesi”. Uluslararası Mühendislik Araştırma ve Geliştirme Dergisi, 10(1), 66-77, 2017.
18. [18] Carvalho-Silva M, Monteiro MTT, de Sá-Soares F, DóriaNóbrega S. “Assessment of forecasting models for patients arrival at Emergency Department”. Operations Research for Health Care, 18, 112-118, 2018.
19. [19] Singhal S, Allen MW, McAnnally JR, Smith KS, Donnelly JP, Wang HE. “National estimates of emergency department visits for pediatric severe sepsis in the United States”. PeerJ, 1, 1-12, 2013.
20. [20] Golmohammadi D. “Predicting hospital admissions to reduce emergency department boarding”. International Journal of Production Economics, 182, 535-544, 2016.
21. [21] Esen H, Kaya Ü. “Bir eğitim araştırma hastanesi acil servis birimine başvuran hasta sayısı tahmini”. Verimlilik Dergisi, 3(3), 129-145, 2021.
22. [22] Lucini FR, Fogliatto FS, da Silveira GJ, Neyeloff JL, Anzanello MJ, Kuchenbecker RS, Schaan, BD. “Text mining approach to predict hospital admissions using early medical records from the emergency department”. International journal of medical informatics, 100, 1-8, 2017.
23. [23] Singer AJ, Thode Jr HC, Peacock IV WF. “Admission rates for emergency department patients with venous thromboembolism and estimation of the proportion of low risk pulmonary embolism patients: a US perspective”. Clinical and experimental emergency medicine, 3(3), 126-131, 2016.
24. [24] Webb BJ, Levin NM, Grisel N, Brown SM, Peltan ID, Spivak ES, Bledsoe J. “Simple scoring tool to estimate risk of hospitalization and mortality in ambulatory and emergency department patients with COVID-19”. Plos one, 17(3), 1-12, 2022.
25. [25] Box G, Jenkins G, Reinsel G, Ljung G. Thrid Edition Time Series Analysis Forecasting and Control. New Jersey, Kanada, John Wiley & Sons, 1995.
26. [26] Toğa G, Atalay B, Toksari MD. “COVID-19 prevalence forecasting using Autoregressive Integrated Moving Average (ARIMA) and Artificial Neural Networks (ANN): Case of Turkey”. Journal Infect Public Health, 14(7), 811-816, 2021.
27. [27] Dagum EB, Giannerini S. “A critical investigation on detrending procedures for non-linear processes”. Journal of Macroeconomics, 28(1), 175-191, 2006.
28. [28] Liu J, Zhao Z, Zhong Y, Zhao C, Zhang G. “Prediction of the dissolved gas concentration in power transformer oil based on SARIMA model”. Energy Reports, 8(5), 1360-1367, 2022.
29. [29] Winters PR. “Forecasting Sales by Exponentially Weighted Moving Averages”. Management science, 6(3), 324-342, 1960.
30. [30] Holt CC. “Forecasting seasonals and trends by exponentially weighted moving averages”. International journal of forecasting, 20(1), 5-10, 2004.
31. [31] Omar MS, Kawamukai H. “Prediction of NDVI using the Holt-Winters model in high and low vegetation regions: A case study of East Africa”. Scientific African, 14, 1-10, 2021.
32. [32] Demir İ, Genç T, Karaboğa HA. “Türkiye Cumhuriyet Merkez Bankası altın rezervinin holt-winters üstel düzleme yöntemi ve yapay sinir ağları ile incelenmesi”. Uluslararası Ekonomi, İşletme ve Politika Dergisi, 2(1), 131-146, 2018.
33. [33] Yang D, Sharma V, Ye ZI, Lim L, Zhao L, Aryaputera AW. “Forecasting of global horizontal irradiance by exponential smoothing, using decompositions”. Energy, 81, 111-119, 2015.
34. [34] C. Voyant, Notton G, Kalogirou S, Nivet ML, Paoli C, Motte Fabrice, Fouilloy A. “Machine learning methods for solar radiation forecasting: A review”. Renewable Energy, 105, 569-582, 2017.
35. [35] Namlı E, Ünlü R, Gül E. “Fiyat tahminlemesinde makine öğrenmesi teknikleri ve doğrusal regresyon yöntemlerinin kıyaslanması; Türkiye’de satılan ikinci el araç fiyatlarının tahminlenmesine yönelik bir vaka çalışması”. Konya Mühendislik Bilimleri Dergisi, 7(4), 806-821, 2019.
36. [36] Januschowski T, Wang Y, Torkkola K, Erkkilä T, Hasson H, Gasthaus J. “Forecasting with trees”. International Journal of Forecasting, 38(4), 1473-1481, 2022.
37. [37] Gradojevic N, Kukolj D, Adcock R, Djakovic V. “Forecasting Bitcoin with technical analysis: A not-so-random forest?”. International Journal of Forecasting, 39(1), 1-17, 2021.
38. [38] Basher SA, Sadorsky P. “Forecasting Bitcoin price direction with random forests: How important are interest rates, inflation, and market volatility?”. Machine Learning with Applications, 9, 1-19, 2022.
39. [39] Torun Y, Ergül Z, Aksöz A, “Optimum enerji verimliliğini hedefleyen rastgele ağaçlar ve yapay arı kolonisi yöntemi ile otonom robotlarda yol planlama algoritması”. Gazi University Journal of Science Part C: Design and Technology, 7(4), 903-915, 2019.
40. [40] Ali J, Khan R, Ahmad N, Maqsood I. “Random forests and decision trees”. International Journal of Computer Science Issues, 9(5), 272-278, 2012.
41. [41] Breiman L. “Random forests”. Machine Learning 2001, 45(1), 5-32, 2001.
42. [42] Caruana R, Niculescu-Mizil A.“An empirical comparison of supervised learning algorithms”. Proceedings of the 23rd İnternational Conference on Machine Learning, Pennsylvania, USA, 25 June 2006.
43. [43] Lewis CD. “Industrial and Business Forecasting Methods: A Practical Guide to Exponential Smoothing and Curve Fitting”. Boston, London, Butterworth Scientific, 1982.
44. [44] Hasan SH, Srivastava P, Talat M. “Biosorption of Pb(II) from water using biomass of Aeromonas hydrophila: Central composite design for optimization of process variables”. Journal of Hazardous Materials, 168(2-3), 1155-1162, 2009.