In this thesis, it is aimed to determine the mortality risk of a patient during or shortly
after the heart surgery by using machine learning techniques based on classification.
The dataset used in this study is obtained from Acıbadem Maslak Hospital. Risk factors
of EuroSCORE which is used to predict the mortality risk of a patient during or shortly
after the heart surgery is used for predicting mortality risk. Because 30-day follow-up
information of patients is not available in the dataset, first the standard EuroSCORE
scores of patients are calculated. Then these risk groups are treated like class labels so
predictions are carried out. Different models are created using Naive Bayes Classifier,
k-Nearest Neighbor Algorithm, Logistic Regression Analysis, ID3 and C4.5 Decision
Tree Algorithms. Performance of the classifiers are compared. Data analysis is carried
out with R language. RStudio was used as a development tool for R codes. Models
derived from Logistic Regression are made available for public via web with Shiny
(shinyapps.io). Another Shiny application is developed for the C4.5 decision tree model
which has best performance.