Predicting Thrombophilia using Neural Networks and Decision Trees

Abstract

The occurrence of thrombophilia during pregnancy results from a complex interaction of inherited and acquired factors, followed by an increase in blood coagulation and subsequent placental ischemic conditions. In this paper, a novel method is presented, whose aim is early identification of the risk of developing thrombophilia in pregnancy. The proposed method is based on machine learning algorithms: decision trees and neural networks. The research uses a dataset consisting of demographic, lifestyle, and clinical information from 35 pregnant women (22 healthy and 13 with thrombophilia). The results show the effectiveness of decision trees and neural networks in accurately predicting the risk of developing thrombophilia in pregnancy. The implications of this research are significant for clinical practice and it provides a valuable tool for early identifying women with high risk of thrombophilia in pregnancy that can enable improvement of preventive measures, such as lifestyle modifications and the use of therapeutic prophylaxis. In conclusion, this paper demonstrates the potential of machine learning algorithms for the prediction of thrombophilia in pregnancy. By combining advanced computational techniques with comprehensive datasets, we can enhance our understanding of thrombophilia in pregnancy risk factors and improve patient outcomes through personalized preventive measures.