Prediction of Epidemic Disease Severity and the Relative Importance of the Factors for Epidemic Disease Using the Machine Learning Methods

Öz

Epidemic diseases have been seen frequently in recent years. Today’s, thanks to advanced database systems, it is possible to reach the clinical and demographic data of citizens. With the help of these data, machine learning algorithms can predict how severe (at home, hospital or intensive care unit) the disease will be experienced by patients in the risk group before the epidemic begins to spread. With these estimates, necessary precautions can be taken. In this study, during the COVID-19 epidemic, the data obtained from the Italian national drug database was used. COVID-19 severity and the features (Age, Diabetes, Hypertension etc.) that affect the severity was estimated using data mining (CRISP-DM method), machine learning approaches (Bagged Trees, XGBoost, Random Forest, SVM) and an algorithm solving the unbalanced class problem (SMOTE). According to the experimental findings, the Bagged Classification and Regression Trees (Bagged CART) yielded higher accuracy COVID-19 severity prediction results than other methods (83.7%). Age, cardiovascular diseases, hypertension, and diabetes were the four highest significant features based on the relative features calculated from the Bagged CART classifier. The proposed method can be implemented without losing time in different epidemic diseases that may arise in the future.

Anahtar Kelimeler

• Epidemic Diseases Severity Prediction
• SARS-COV-2
• Machine Learning
• Data Mining
• Random Forest
• Extreme Gradient Boosting

Makine Öğrenimi Yöntemlerini Kullanarak Salgın Hastalık Şiddetinin ve Salgın Hastalık Faktörlerinin Göreceli Önemlerinin Tahmin Edilmesi

Öz

Salgın hastalıklar son yıllarda sıklıkla görülmektedir. Günümüzde gelişmiş veritabanı sistemleri sayesinde vatandaşların klinik ve demografik verilerine ulaşmak mümkündür. Bu veriler yardımıyla makine öğrenme algoritmaları, salgın yayılmaya başlamadan önce risk grubundaki hastaların hastalığın ne kadar şiddetli (evde, hastanede veya yoğun bakım ünitesinde) yaşayacağını tahmin edebilir. Bu tahminler ile gerekli önlemler alınabilir. Bu çalışmada, COVID-19 salgını sırasında İtalya ulusal ilaç veri tabanından elde edilen veriler kullanılmıştır. COVID-19 şiddeti ve şiddeti etkileyen özellikler (Yaş, Diyabet, Hipertansiyon vb.), veri madenciliği (CRISP-DM Metodu), makine öğrenmesi yaklaşımları (Bagged Trees, XGBoost, Random Forest, SVM) ve dengesiz sınıf problemini çözen bir algoritma (SMOTE) kullanılarak tahmin edilmiştir. Deneysel bulgulara göre Torbalı Sınıflandırma ve Regresyon Ağaçları (Bagged CART), diğer yöntemlere göre (%83,7) daha yüksek doğrulukta COVID-19 şiddeti tahmin sonuçları vermiştir. Torbalı CART sınıflandırıcısından hesaplanan göreli özelliklere dayalı olarak yaş, kardiyovasküler hastalıklar, hipertansiyon ve diyabet en önemli dört özellik olarak tahmin edilmiştir. Önerilen yöntem ileride ortaya çıkabilecek farklı salgın hastalıklarda zaman kaybetmeden uygulanabilecektir.

Kaynakça

1. Işık A. SALGIN EKONOMİSİNE GENEL BİR BAKIŞ. Int Anatolia Acad Online J [Internet]. 2021;7(2). Available from: https://dergipark.org.tr/en/download/article-file/1933517
2. Pandemi [Internet]. 2022. Available from: https://tr.wikipedia.org/wiki/Pandemi
3. Olgun Eker E. Effects Of Climate Change On Health. 2020;13–23.
4. Bhadoria P, Gupta G, Agarwal A. Viral pandemics in the past two decades: An overview. J Fam Med Prim Care [Internet]. 2021;10(8):2745. Available from: https://journals.lww.com/jfmpc/Fulltext/2021/10080/Viral_Pandemics_in_the_Past_Two_Decades__An.5.aspx
5. Ming-Syan Chen, Jiawei Han, Yu PS. Data mining: an overview from a database perspective. IEEE Trans Knowl Data Eng [Internet]. 1996;8(6):866–83. Available from: http://ieeexplore.ieee.org/document/553155/
6. KARTAL E, BALABAN ME, BAYRAKTAR B. KÜRESEL COVID-19 SALGINININ DÜNYADA VE TÜRKİYE’DE DEĞİŞEN DURUMU VE KÜMELEME ANALİZİ. İstanbul Tıp Fakültesi Derg [Internet]. 2021 Jan 20;84(1). Available from: https://iupress.istanbul.edu.tr/tr/journal/jmed/article/kuresel-covid-19-salgininin-dunyada-ve-turkiyede-degisen-durumu-ve-kumeleme-analizi
7. Komenda M, Bulhart V, Karolyi M, Jarkovský J, Mužík J, Májek O, et al. Complex Reporting of the COVID-19 Epidemic in the Czech Republic: Use of an Interactive Web-Based App in Practice. J Med Internet Res [Internet]. 2020 May 27;22(5):e19367. Available from: http://www.jmir.org/2020/5/e19367/
8. Rivai MA. Analysis of Corona Virus spread uses the CRISP-DM as a Framework: Predictive Modelling. Int J Adv Trends Comput Sci Eng [Internet]. 2020 Jun 25;9(3):,2987-2994. Available from: http://www.warse.org/IJATCSE/static/pdf/file/ijatcse76932020.pdf

9. UTAMA ID, SUDIRMAN ID. OPTIMIZING DECISION TREE CRITERIA TO IDENTIFY THE RELEASED FACTORS OF COVID-19 PATIENTS IN SOUTH KOREA. J Theor Appl Inf Technol. 2020;98(16):3305–15.
10. Jaggia S, Kelly A, Lertwachara K, Chen L. Applying the CRISP‐DM Framework for Teaching Business Analytics. Decis Sci J Innov Educ [Internet]. 2020 Oct 21;18(4):612–34. Available from: https://onlinelibrary.wiley.com/doi/10.1111/dsji.12222
11. John M, Shaiba H. Main factors influencing recovery in MERS Co-V patients using machine learning. J Infect Public Health [Internet]. 2019 Sep;12(5):700–4. Available from: https://linkinghub.elsevier.com/retrieve/pii/S1876034119301297
12. Forna A, Nouvellet P, Dorigatti I, Donnelly CA. Case Fatality Ratio Estimates for the 2013–2016 West African Ebola Epidemic: Application of Boosted Regression Trees for Imputation. Clin Infect Dis [Internet]. 2020 Jun 10;70(12):2476–83. Available from: https://academic.oup.com/cid/article/70/12/2476/5536742
13. Colubri A, Hartley MA, Siakor M, Wolfman V, Felix A, Sesay T, et al. Machine-learning Prognostic Models from the 2014–16 Ebola Outbreak: Data-harmonization Challenges, Validation Strategies, and mHealth Applications. EClinicalMedicine. 2019;11:54–64.
14. Hu C-A, Chen C-M, Fang Y-C, Liang S-J, Wang H-C, Fang W-F, et al. Using a machine learning approach to predict mortality in critically ill influenza patients: a cross-sectional retrospective multicentre study in Taiwan. BMJ Open [Internet]. 2020 Feb 25;10(2):e033898. Available from: https://bmjopen.bmj.com/lookup/doi/10.1136/bmjopen-2019-033898
15. Patel SJ, Chamberlain DB, Chamberlain JM. A Machine Learning Approach to Predicting Need for Hospitalization for Pediatric Asthma Exacerbation at the Time of Emergency Department Triage. Cloutier R, editor. Acad Emerg Med [Internet]. 2018 Dec 29;25(12):1463–70. Available from: https://onlinelibrary.wiley.com/doi/10.1111/acem.13655
16. Ahamad MM, Aktar S, Rashed-Al-Mahfuz M, Uddin S, Liò P, Xu H, et al. A machine learning model to identify early stage symptoms of SARS-Cov-2 infected patients. Expert Syst Appl [Internet]. 2020 Dec;160:113661. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0957417420304851
17. Banerjee A, Ray S, Vorselaars B, Kitson J, Mamalakis M, Weeks S, et al. Use of Machine Learning and Artificial Intelligence to predict SARS-CoV-2 infection from Full Blood Counts in a population. Int Immunopharmacol [Internet]. 2020 Sep;86:106705. Available from: https://linkinghub.elsevier.com/retrieve/pii/S1567576920315770
18. Malki Z, Atlam E-S, Hassanien AE, Dagnew G, Elhosseini MA, Gad I. Association between weather data and COVID-19 pandemic predicting mortality rate: Machine learning approaches. Chaos, Solitons & Fractals [Internet]. 2020 Sep;138:110137. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0960077920305336
19. García-Ordás MT, Arias N, Benavides C, García-Olalla O, Benítez-Andrades JA. Evaluation of Country Dietary Habits Using Machine Learning Techniques in Relation to Deaths from COVID-19. Healthcare [Internet]. 2020 Sep 29;8(4):371. Available from: https://www.mdpi.com/2227-9032/8/4/371
20. Kivrak M, Guldogan E, Colak C. Prediction of death status on the course of treatment in SARS-COV-2 patients with deep learning and machine learning methods. Comput Methods Programs Biomed [Internet]. 2021 Apr;201:105951. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0169260721000250
21. Schröer C, Kruse F, Gómez JM. A Systematic Literature Review on Applying CRISP-DM Process Model. Procedia Comput Sci [Internet]. 2021;181:526–34. Available from: https://linkinghub.elsevier.com/retrieve/pii/S1877050921002416
22. Arslan, A. K. & Çolak, C. BKSY: Bilgi Keşfi Süreci Yazılımı [Web-tabanlı yazılım] biostatapps.inonu.edu.tr [Internet]. Available from: http://biostatapps.inonu.edu.tr/BKSY/
23. Bravi F, Flacco ME, Carradori T, Volta CA, Cosenza G, De Togni A, et al. Predictors of severe or lethal COVID-19, including Angiotensin Converting Enzyme inhibitors and Angiotensin II Receptor Blockers, in a sample of infected Italian citizens. Shimosawa T, editor. PLoS One [Internet]. 2020 Jun 24;15(6):e0235248. Available from: https://dx.plos.org/10.1371/journal.pone.0235248
24. Menardi G, Torelli N. Training and assessing classification rules with imbalanced data. Data Min Knowl Discov [Internet]. 2014 Jan 30;28(1):92–122. Available from: http://link.springer.com/10.1007/s10618-012-0295-5
25. Turlapati VPK, Prusty MR. Outlier-SMOTE: A refined oversampling technique for improved detection of COVID-19. Intell Med [Internet]. 2020 Dec;3–4:100023. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2666521220300235
26. Starling JK, Mastrangelo C, Choe Y. Improving Weibull distribution estimation for generalized Type I censored data using modified SMOTE. Reliab Eng Syst Saf [Internet]. 2021 Feb;107505. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0951832021000661
27. Chawla N V., Bowyer KW, Hall LO, Kegelmeyer WP. SMOTE: Synthetic Minority Over-sampling Technique. 2011 Jun 9; Available from: http://arxiv.org/abs/1106.1813
28. Haibo He, Yang Bai, Garcia EA, Shutao Li. ADASYN: Adaptive synthetic sampling approach for imbalanced learning. In: 2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence) [Internet]. IEEE; 2008. p. 1322–8. Available from: http://ieeexplore.ieee.org/document/4633969/
29. Pavlov YL. Random forests. Random For. 2019;1–122.
30. Chen T, Guestrin C. XGBoost. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining [Internet]. New York, NY, USA: ACM; 2016. p. 785–94. Available from: https://dl.acm.org/doi/10.1145/2939672.2939785
31. Weston J, Mukherjee S, Chapelle O, Pontil M, Poggio T, Vapnik V. Feature selection for SVMs. Adv Neural Inf Process Syst. 2001;
32. Colak C, Colak MC, Ermis N, Erdil N, Ozdemir R. Prediction of cholesterol level in patients with myocardial infarction based on medical data mining methods. Kuwait J Sci [Internet]. 2016;43(Vol. 43 No. 3 (2016): Kuwait Journal of Science):86–90. Available from: https://journalskuwait.org/kjs/index.php/KJS/article/view/875/139
33. Praagman J. Classification and regression trees. Eur J Oper Res [Internet]. 1985 Jan;19(1):144. Available from: https://linkinghub.elsevier.com/retrieve/pii/0377221785903212
34. Islam MM, Rahman MJ, Chandra Roy D, Maniruzzaman M. Automated detection and classification of diabetes disease based on Bangladesh demographic and health survey data, 2011 using machine learning approach. Diabetes Metab Syndr Clin Res Rev [Internet]. 2020 May;14(3):217–9. Available from: https://linkinghub.elsevier.com/retrieve/pii/S1871402120300448
35. Arlot S, Celisse A. A survey of cross-validation procedures for model selection. Stat Surv [Internet]. 2010;4:40–79. Available from: http://projecteuclid.org/euclid.ssu/1268143839
36. YAŞAR Ş, ARSLAN A, ÇOLAK C, YOLOĞLU S. A Developed Interactive Web Application for Statistical Analysis: Statistical Analysis Software. Middle Black Sea J Heal Sci [Internet]. 2020 Aug 31;226–38. Available from: https://dergipark.org.tr/tr/doi/10.19127/mbsjohs.704456
37. Wang K-J, Adrian AM, Chen K-H, Wang K-M. A hybrid classifier combining Borderline-SMOTE with AIRS algorithm for estimating brain metastasis from lung cancer: A case study in Taiwan. Comput Methods Programs Biomed [Internet]. 2015 Apr;119(2):63–76. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0169260715000577
38. Koziarski M. Radial-Based Undersampling for imbalanced data classification. Pattern Recognit [Internet]. 2020 Jun;102:107262. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0031320320300674
39. Zhu Z, Wang Z, Li D, Du W. NearCount: Selecting critical instances based on the cited counts of nearest neighbors. Knowledge-Based Syst [Internet]. 2020 Feb;190:105196. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0950705119305325
40. Liu B, Tsoumakas G. Dealing with class imbalance in classifier chains via random undersampling. Knowledge-Based Syst [Internet]. 2020 Mar;192:105292. Available from: https://linkinghub.elsevier.com/retrieve/pii/S0950705119305830
41. YAVAŞ M, GÜRAN A, UYSAL M. Covid-19 Veri Kümesinin SMOTE Tabanlı Örnekleme Yöntemi Uygulanarak Sınıflandırılması. Eur J Sci Technol [Internet]. 2020 Aug 15;258–64. Available from: https://dergipark.org.tr/tr/doi/10.31590/ejosat.779952
42. Guner R, Hasanoglu I, Kayaaslan B, Aypak A, Akinci E, Bodur H, et al. Comparing ICU admission rates of mild/moderate COVID-19 patients treated with hydroxychloroquine, favipiravir, and hydroxychloroquine plus favipiravir. J Infect Public Health [Internet]. 2021 Mar;14(3):365–70. Available from: https://linkinghub.elsevier.com/retrieve/pii/S1876034120307735
43. Rohila VS, Gupta N, Kaul A, Sharma DK. Deep Learning Assisted COVID-19 Detection using full CT-scans. Internet of Things [Internet]. 2021 Feb;100377. Available from: https://linkinghub.elsevier.com/retrieve/pii/S2542660521000214