Title: Machine learning techniques for dental disease prediction

Abstract

Oral diseases are increasing at the same rate as infectious diseases and non communicable diseases all over the world. More than eighty percent of the total population suffers from one or more dental diseases, of which periodontitis, gingivitis, and carcinoma are among them. In this work, we used a machine learning approach for dental disease prediction in the context of the daily behavior of the people of a country. We discussed with the concerned doctors and the dentist the important factors of dental disease. By talking to a doctor or someone with experience in the field, we can find out if there are certain habits or characteristics that cause tooth loss or are more likely to cause disease. With all these important factors in mind, we started collecting data from real people and dental disease patients. Generally, processing the data is very important in creating a machine learning model because, while training from the collected original data, some unexpected or incorrect data may show bias during model training. After data collection and preprocessing, we used nine eminent machine-learning algorithms namely k-nearest neighbors, logistic regression, SVM, naïve Bayes, classification and regression trees, random forest, multilayer perception, adaptive boosting, and linear discriminant analysis. For the task of assessment, we reviewed the performance of each classifier using accuracy and some noteworthy performance metrics. The logistic regression classifier outflanks every single other classifier regarding all measurements utilized by accomplishing an accuracy approaching 95.89%. On the basis thereof, AdaBoost shows not only deficient consequences of an accuracy approaching 34.69% but also some deficient outcomes in other metrics.