Prediction of pediatric HIV/Aids survival in Nigeria using naïve Bayes’ approach

Abstract

Epidemic diseases have highly destructive effects around the world and these diseases have affected both developed and developing nations. Disease epidemics are common in developing nations especially in Sub Saharan Africa in which Human Immunodeficiency Virus /Acquired Immunodeficiency Disease Syndrome (HIV/AIDS) is the most serious of all. This paper presents a prediction of pediatric HIV/AIDS survival in Nigeria. Data are collected from 216 pediatric HIV/AIDS patients who were receiving antiretroviral drug treatment in Nigeria was used to develop a predictive model for HIV/AIDS survival based on identified variables. Interviews were conducted with the virologists and pediatricians to identify the variables predictive for HIV/AIDS survival in pediatric patients. 10-fold cross validation method was used in performing the stratification of the datasets collected into training and testing datasets following data preprocessing of the collected datasets. The model was formulated using the naïve Bayes’ classifier – a supervised machine learning algorithm based on Bayes’ theory of conditional probability and simulated on the Waikato Environment for Knowledge Analysis [WEKA] using the identified variables, namely: CD4 count, viral load, opportunistic infection and the nutrition status of the pediatric patients involved in the study. The results showed 81.02% accuracy in the performance of the naïve Bayes’ classifier used in developing the predictive model for HIV/AIDS survival in pediatric patients. In addition, the area under the receiver operating characteristics [ROC] curve had a value of 0.933 which showed how well the developed predictive model was able to discriminate between survived and non-survived cases. Model validation was performed by comparing the model results with that of historical data from two (2) selected tertiary institutions in Nigeria.

Keywords: HIV/AIDS survival, naïve Bayes’ classifier, predictive model, pediatrics, machine learning