Application of Machine Learning For Prediction Dental Material Wear

Resin composites are commonly applied as the material for dental restoration. Wear of these materials is a major issue. In this study specimens made of dental composite materials were subjected to an in-vitro test in a pin-on-disc tribometer. Four different dental composite materials applied in the experiment were soaked in a solution of chewing tobacco for certain days before being removed and put through a wear test. Subsequently, four different machine learning (ML) algorithms (AdaBoost, CatBoost, Gradient Boosting, Random Forest) were implemented for developing models for the prediction of wear of dental materials. AdaBoost, CatBoost, Gradient Boosting and Random Forest model show an MAE of 0.7011, 0.0773, 0.0771 and 0.2199. AdaBoost model performs poorly in comparison to other models.