Presenting a Model for Periodontal Disease Diagnosis Using Two Artificial Neural Network Algorithms

Abstract

Background: Artificial neural networks (ANNs) can be used in various medical cases due to their high performance in learning the relationship between variables. Periodontal diseases are common oral infectious diseases that can cause tooth loss, if not treated. Objectives: The current study aimed at evaluating the role of ANNs in periodontal disease diagnosis. Methods: The data were collected from 190 periodontal disease cases in Zahedan dentistry school from 2015 to 2016. Five variables including age, gender, plaque index, probing pocket depth, and clinical attachment loss index were evaluated. The patients were divided into two groups of training (n = 160), and testing (n = 30). In the current study model, two Levenberg-Marquardet (LM) and scaled conjugate gradient (SCG) algorithms were used, and the results were compared in terms of the number of iterations and the mean square error (MSE). Results: The obtained results showed that the LM algorithm with fewer iterations and a minimum MSE, had a better performance than the SCG algorithm. Conclusions: ANNs can be used with low error as an effective tool to diagnose periodontal diseases.