Abstract

OLAYEMI, O. C., OLASEHINDE O. O., OJOKOH, B. A., PETER, A. I

Lower Respiratory Tract Infections (LRTIs) are the major causes of mortality in paediatrics. Literature reviews reveal that LRTIs accounted for more than a million children morbidity and mortality yearly due to a lack of prompt diagnosis or no diagnosis due to a shortage of medical experts and medical facilities. The use of Machine learning (ML) techniques can be employed to fill this gap. This study evaluates ML models for a prompt and timely diagnosis of LRTIs in developing countries. The LRTIs dataset used in this study was obtained from The Federal Medical Centre, Owo, Nigeria. Relevant features of the dataset based on Information gain and Correlation feature selection techniques were used to build five machine learning predictive models; Naïve Bayes (NB), Multi-Layer Perceptron (MLP), K-Nearest Neighbor (KNN), Support Vector Machine (SVM), and Random Forest (RF). The predictive models' evaluation was based on standard performance metrics (accuracy, sensitivity, specificity, and precision).The experimental results show that the information gain predictive models perform better than the correlation predictive models. The RF predictive model of the Information Gain feature selection method recorded the best accuracy of 98.53%. RF is therefore recommended for making decisions concerning respiratory infection diagnosis.

PDF

Share this article

Get the App