Makine öğrenimi yöntemleri ile kalp krizinin sınıflandırılması ve ilişkili risk faktörlerinin belirlenmesi için bir model oluşturulması

Abstract

Amaç: Kalp krizlerinde (KK) hayatta kalma oranlarının artması, erken müdahale ve tedaviye bağlıdır. Bu çalışmada, makine öğrenmesi yöntemlerinden biri olan Stokastik Gradient Boosting (SGB) yöntemi kullanılarak KK ile ilişkili olabilecek faktörlerin tahmin edilmesi ve hangi faktörün daha etkili olduğunun belirlenmesi amaçlanmaktadır. Yöntemler: Araştırmada açık erişimli veri seti kullanıldı. Modellemede 5 katlı çapraz doğrulama yöntemi kullanılmış ve veri seti %80:%20 olacak şekilde eğitim ve test veri setlerine bölünmüştür. Model değerlendirmesi için doğruluk (ACC), dengeli doğruluk (b-ACC), duyarlılık (SE), özgüllük (SP), pozitif tahmin değeri (ppv), negatif tahmin değeri (npv) ve F1 skoru metrikleri kullanıldı. Bulgular: Modelleme ile performans metriklerinden elde edilen sonuçlar ACC, b-ACC, SE, SP, ppv, npv, F1 puanı çin %98,9, %98,7, %99,4, %98,0, %98,8, %99 ve %99,1 olmuştur. Değişken önem değerlerine göre sırasıyla troponin ve CK-MB'nin KK ile ilişkili olduğu görülmektedir. Sonuç: Modelleme sonuçlarına göre kalp kriziyle ilişkili olabilecek faktörler makine öğrenmesi yöntemiyle yüksek doğrulukla belirlendi. Bu iki enzim sayesinde kalp krizi geçirme riski taşıyan bireylerde erken tanı yapılabilmekte, kötü gidişat ve ölümlerin önüne geçilebilmektedir.

Keywords

• Kalp krizi
• sınıflandırma
• makine öğrenmesi
• risk faktörü

Establishing a Model for the Classification of Heart Attack and Identification of Associated Risk Factors with Machine Learning Methods

Abstract

Object: Increased survival rates in heart attacks (HAs) depend on early intervention and treatment. In this study, it is aimed to predict the factors that may be associated with HA and to determine which factor is more effective by using Stochastic Gradient Boosting (SGB) method, one of the machine learning methods. Methods: An open access data set was used in the study. The 5-fold cross-validation method was used in modeling and the data set was divided into training and test data sets as 80%:20%. Accuracy (ACC), balanced accuracy (b-ACC), sensitivity (SE), specificity (SP), positive predictive value (ppv), negative predictive value (npv) and F1 score metrics were used for model evaluation. Results: The results obtained from the performance metrics with the modeling were 98.9%, 98.7%, 99.4%, 98.0%, 98.8%, 99%, and 99.1% for ACC, b-ACC, SE, SP, ppv, npv, and F1-score, respectively. According to variable importance values, troponin and CK-MB appear to be associated with HA, respectively. Conclusion: According to the modeling results, factors that may be associated with heart attack were determined with high accuracy by machine learning method. Thanks to these two enzymes, early diagnosis can be made in individuals at risk of having a heart attack, and poor prognosis and deaths can be prevented.

Keywords

• Heart attack
• classification
• machine learning
• risk factor

References

1. Liu Z, Meng D, Su G, Hu P, Song B, Wang Y, et al. Ultrafast Early Warning of Heart Attacks through Plasmon-Enhanced Raman Spectroscopy using Collapsible Nanofingers and Machine Learning. Small (Weinheim an der Bergstrasse, Germany). 2023;19(2):e2204719.
2. Maghdid SS, Rashid T, Ahmed S, Zaman K, Rabbani M. Analysis and prediction of heart attacks based on design of intelligent systems. J Mech Contin Math Sci. 2019;14(4):628-45.
3. Organization WH. Cardiovascular diseases (cvds). http://www who int/mediacentre/factsheets/fs317/en/index html. 2009.
4. https://www.nhs.uk/conditions/heart-attack/diagnosis/#:~:text=An%20ECG%20machine%20records%20these,about%205%20minutes%20to%20do. [cited 2023 07.04].
5. Arslankaya S, Çelik MT. Prediction of heart attack using fuzzy logic method and determination of factors affecting heart attacks. International Journal of Computational and Experimental Science and Engineering. 2021;7(1):1-8.
6. Zimetbaum PJ, Josephson ME. Use of the electrocardiogram in acute myocardial infarction. New England Journal of Medicine. 2003;348(10):933-40.
7. Gupta V, Mittal M. A novel method of cardiac arrhythmia detection in electrocardiogram signal. International Journal of Medical Engineering and Informatics. 2020;12(5):489-99.
8. Leung DY, Davidson PM, Cranney GB, Walsh WF. Thromboembolic risks of left atrial thrombus detected by transesophageal echocardiogram. The American journal of cardiology. 1997;79(5):626-9.

9. Anshori M, Haris MS. Predicting Heart Disease using Logistic Regression. Knowledge Engineering and Data Science. 2022;5(2):188-96.
10. Friedman JH. Stochastic gradient boosting. Computational statistics & data analysis. 2002;38(4):367-78.
11. Ye J, Chow J-H, Chen J, Zheng Z, editors. Stochastic gradient boosted distributed decision trees. Proceedings of the 18th ACM conference on Information and knowledge management; 2009.
12. Lawrence R, Bunn A, Powell S, Zambon M. Classification of remotely sensed imagery using stochastic gradient boosting as a refinement of classification tree analysis. Remote sensing of environment. 2004;90(3):331-6.
13. Zeynep T, Çiçek İB, Güldoğan E. Performance evaluation of the deep learning models in the classification of heart attack and determination of related factors. The Journal of Cognitive Systems. 2020;5(2):99-103.
14. Halıcı Z, Yasin Bayır HS, Çadırcı E, Keleş MS, Bayram E. Investigation of the Effects of Amiodarone on Serum Eritropoietin Levels in İsoproterenol-induced Acute and Chronic Myocardial Infarct Model of Rats. Avrasya J Med, 2006;38(3): 68-72.
15. Storrow AB, Gibler WB. Chest pain centers: diagnosis of acute coronary syndromes. Annals of emergency medicine. 2000;35(5):449-61.
16. Roger VL. Epidemiology of myocardial infarction. The Medical clinics of North America. 2007;91(4):537-52; ix.
17. Özdemir G, Bilen Ö, Ateş SC. Establishment of a Decision Support System for Determining the Risk Probability of Heart Attack in Hospital Emergency Visitors. Düzce Üniv Bilim ve Tek Der. 2022;10(4):2093-106.
18. Yöntem M, Erdoğdu BS, Akdoğan M, Kaleli S. The Importance of Cardiac Markers in Diagnosis of Acute Myocardial Infarction. Online Türk Sağlık Bilimleri Dergisi. Online Türk Sağlık Bilimleri Der. 2017;2(4):11-7.
19. Lewandrowski K, Chen A, Januzzi J. Cardiac markers for myocardial infarction: a brief review. Pathology Patterns Reviews. 2002;118(suppl_1):S93-S9.
20. Filatov V, Katrukha A, Bulargina T, Gusev N. Troponin: structure, properties, and mechanism of functioning. Biochemistry c/c of Biokhimiia. 1999;64:969-85.