Yoğun Bakım Ünitelerinde Yatan Hastalara İlişkin Mortalite ve Yatış Süresine Etki Eden Faktörlerin Veri Madenciliği Yöntemleriyle İncelenmesi,Yoğun Bakım Ünitelerinde Yatan Hastalara İlişkin Mortalite ve Yatış Süresine Etki Eden Faktörlerin Veri Madenciliği Yöntemleriyle İncelenmesi
xmlui.mirage2.itemSummaryView.MetaDataShow full item record
In this thesis, association rules, classification and regression models of data mining methods were used in order to estimate mortality and length of stay in intensive care units. A total of seven classification models, C4.5, CART, random forest, logistic regression, support vector machine, artificial neural network and naive bayes, were built using a total of 4,233 patient data for estimating mortality. Random forest model, which has the best predictive performance, was compared with APACHE II mortality prediction model which is commonly used in intensive care units as a logistic regression based model, and it was found that random forest model performed better. In order to predict length of stay in intensive care units; linear regression, support vector machines, random forest and artificial neural network models were built and artificial neural network performed better then other models. In addition, variable significance was found out from classification and regression models and as a result, most important factors affecting mortality and length of stay were determined.
The following license files are associated with this item: