What question did this study set out to answer?

Assess the predictive value of machine learning models for dental implant failure.

March 21, 2026

Predicting dental implant failure using machine learning: comparative evaluation of Random Forest, gradient boosting, and logistic regression with feature importance analysis

Key Points

Assess the predictive value of machine learning models for dental implant failure.
Evaluated logistic regression, Random Forest, and gradient boosting models.
Used an open-access dataset including demographic, surgical, prosthetic, and systemic variables.
Conducted feature importance analysis to identify key predictors.
Random Forest achieved 85% accuracy, 0.79 ROC-AUC, and an F1 score of 0.92.
Gradient boosting performed slightly less effectively than Random Forest.
Logistic regression showed lower sensitivity compared to ensemble models.

Abstract

Dental implant failure is influenced by anatomical, systemic, and procedural factors. This study assessed the predictive value of machine learning models: logistic regression, Random Forest, and gradient boosting, using an open-access dataset (Liu et al. 2018) containing demographic, surgical, prosthetic, and systemic variables. Random Forest performed best (accuracy 0.85, ROC-AUC 0.79, F1 = 0.92, recall 0.97), followed by gradient boosting, while logistic regression showed lower sensitivity. Feature importance analysis identified implant location, sinus augmentation, implant dimensions, and patient age as key predictors. Ensemble models and interpretable feature metrics demonstrate strong potential for improving clinical risk stratification in implant dentistry.

Perguntar à IA

Bookmark