Predicting Survival HIV/AIDS Patients Using Machine Learning Methods: Artificial Neural Network and Decision Tree Models
Ladislas NDABARUSHIMANA, Papa Ibrahima Ngom, Paul Python Ndekou, Abdou Kâ Diongue, Emmanuel Banzubaze · 2023
In Burundi, Anti Retroviral Therapy(ART) helps to control infected people living with Immunodeficiency Virus/Acquired Immune Deficiency Syndrome(HIV/AIDS), but the rate of deaths is still high. This paper presents a model for predicting the survival of HIV/AIDS patients treated in health institution” Association Nationale de Soutien aux Seropositifs et Malades du SIDA” (ANSS) in Burundi. Algorithms of Artificial Neural Network(ANN) and Decision Tree(DT) were used for predicting the survival of People Living With HIV(PLWHIV) under the ANSS health care departments. The results show that the average probability of survival of patients at ANSS institution was 90.2% for an average follow-up period of 5 years. The developed model will help ANSS institution and the Ministry of Health to define effective management strategies for these HIV/AIDS patients.