Keras Slim ResNet-Based Prediction of Furcation Management Recommendations Among Dentists.

Purpose: Furcation involvement occurs when periodontal deterioration reaches the roots of a multirooted tooth, making diagnosis and treatment difficult. Furcation management has shifted from periodontal care to unnecessary extraction and prosthetic replacement with or without implants. Prediction modeling with cutting-edge machine learning algorithms was used to select treatments from questionnaire responses. We used Keras ResNet to forecast dentists' furcation management recommendations. Materials and Methods: The participants were dentists with undergraduate or postgraduate degrees in fields other than periodontics and 5 years of practical experience. The study comprised 437 South Indian dentists aged 28-60 years. The author estimated the sample size based on previous research. We compared findings from a data robot tool employing a state-of-the-art model with Keras Slim and Light Gradient Boosting. Data were split 80/20 between training and testing. Results: Keras ResNet and light gradient improved accuracy by 84% in predicting the target class of furcation-related tooth treatment suggestions. Conclusion: The Keras Slim ResNet-based questionnaire-prediction model among dentists has demonstrated good accuracy, helping predict referral patterns for managing furcation-involved teeth. In addition, it encourages general dentists to refer complex furcation cases to periodontists for expert care consistently.