Machine Learning and Feature Selection Based Cardiovascular Disease Prediction

SÜZGEN E. E., GÜNGÖR ULUTAŞ E., ŞAHİN M.

3rd International Congress of Electrical and Computer Engineering, ICECENG 2024, Bandirma, Türkiye, 27 - 30 Kasım 2024, ss.223-236, (Tam Metin Bildiri) identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1007/978-3-031-88999-8_17
  • Basıldığı Şehir: Bandirma
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.223-236
  • Anahtar Kelimeler: Classification, Feature selection, Heart disease, Machine learning, WOA
  • Yozgat Bozok Üniversitesi Adresli: Evet

Özet

Cardiovascular diseases (CVD) continue to be the primary cause of death globally, highlighting the need for innovative approaches in early diagnosis and risk assessment. This study explores the application of machine learning techniques to enhance the prediction accuracy of heart disease, utilizing the Z-Alizadeh Sani dataset, which includes 303 patients characterized by 54 features. We initially applied six machine learning algorithms: Logistic Regression, Random Forest, XGBoost, AdaBoost, Gradient Boosting, and LightGBM, to evaluate their effectiveness in predicting CVD. To optimize the model performance, the Whale Optimization Algorithm (WOA) was employed for feature selection, refining the input features to enhance classification accuracy. After applying feature selection using the Whale Optimization Algorithm (WOA), model performances varied slightly, with LightGBM achieving the highest accuracy of 0.8514 (F1-Score of 0.7182), while Gradient Boosting and Random Forest maintained competitive accuracies of 0.8482 and 0.8446, respectively. This dual-phase experimental approach underscored the impact of strategic feature selection on model performance, highlighting the potential of machine learning models combined with optimization algorithms in providing non-invasive and reliable heart disease prediction. Our findings support the advancement of effective decision support systems to aid healthcare professionals in diagnosing and managing cardiovascular diseases.