SUPERVISED LEARNING TECHNIQUE FOR PREDICTING PROGNOSTIC RISK OF MALARIA AND TYPHOID FEVER

Abstract

Malaria and typhoid are serious infectious diseases of the digestive system, caused by ingesting food or water 
contaminated with the bacillus salmonella. The rate at which malaria and typhoid fever kills particularly in West 
Africa is alarming, and thereby require solution development and knowledge-based decision support. At times, 
rarely available facilities could be occasioned by inadequate medical personnel among other germane issues that 
require urgent redress. In this study, supervised learning technique was adopted to improve the prognostic prediction 
of malaria and typhoid fever. Improved predictive system for malaria and typhoid prognosis was developed using 
Naïve Bayes model, and experimented with localized clinical dataset using Visual Studio.net and ActiveX Data 
Object. Naive Bayes approach was adopted as supervised learning to sequence of symptoms for every observation; 
by analyzing parameter values into a particular diagnostic label being the likely predictor in a training set. 
Experimental evaluation of the system was measured; giving seventy-two percent (72%) accuracy and eighty-one 
percent (81%) precision as performance metrics.