Development of a Model for Recurrent Tonsillitis in Paediatric Patient

Abstract

This study identifies the risk factors of recurrent tonsillitis in pediatric patient which in turn are the variables used in developing a predictive model for predicting the risk of recurrent tonsillitis. This is achieved by eliciting knowledge on the risk factors of recurrent tonsillitis, formulating the model using the variables and simulating the model using MATLAB tool. Interviews were conducted with the pediatrician and existing literature was studied on the knowledge of study in order to identify the variables for recurrent tonsillitis. Seven (7) data from tonsillitis patients were collected from Wesley Guild Hospital, Ilesha. Predictive model was formulated using the fuzzy logic model and simulated on MATLAB R2016a. Fuzzy logic was used as the predictive model to determine the risk of recurrent tonsillitis. The stages involved in the process are four (4) which includes: fuzzification, rule production, aggregation and defuzzification. The identified variables were given crisp values and within a membership function of 0 and 1. The simulated result of the fuzzy logic model was done using MATLAB which involved formulation of the fuzzy logic inference system (FIS) which was carried out by the MATLAB tool. The variables which are the risk factors were used to build the fuzzy logic inference system (FIS) to determine the risk of recurrent tonsillitis. Possible combinations of rules were given for the variables and the rules were used in the inference engine to predict the output of the model whether it is no, low, moderate or high risk of recurrent tonsillitis. The validation was done on the data gotten from Wesley Guild Hospital Ilesha from 7 patients. In conclusion, out of the seven (7) patients test data provided, five (5) patients have low risk, two (2) patients have moderate risk, no patients have no low risk and no patients have high risk of recurrent tonsillitis with 100% test accuracy.