Prediction Model for Diagnosis of Pulpitis Diseases using Learning Vector Quantization 3

Poverty and mobility limitations are among the main factors that hinder people from regular dental visits. Hence, many people, especially from remote areas, did not get the required education on proper dental hygiene and early detection and treatment for dental disease. Pulpitis is one of the frequent dental diseases. A free online diagnosis of Pulpitis would be helpful to users who are having difficulties visiting a dentist. However, the challenge in developing an online advisor or expert system is to create a high accuracy prediction model. One method proven effective in building the required classification model is an Artificial Neural Network (ANN). In this research, we developed the Pulpitis disease prediction model using the LVQ3 algorithm. This developed model can classify five classes of Pulpitis diseases based on 13 symptoms. In addition, we also conducted experimental testing with eight learning rates, eight windows, a maximum of 100 epochs, and time is taken parameters to get the highest accuracy modelling. Based on experimental testing, LVQ3 obtained an average accuracy of 97.5% on training data allocation of 80%. In terms of time taken, the system using the LVQ3 algorithm requiring a total processing time of 328 minutes 23 seconds, with an average for one processing time is 1 minute 42 seconds. Therefore, based on the test results, we concluded that this web-based prediction system has the potential to be used as a solution for the community to get the Pulpitis diseases diagnosis as early as possible.