Facial biotype classification for orthodontic treatment planning using an alternative learning algorithm for tree...

Area of Science:

• Orthodontics
• Biometrics
• Machine Learning

Background:

• Facial growth significantly impacts orthodontic treatment planning for children and adolescents.
• Accurate facial biotype classification is essential for determining appropriate treatment mechanics.
• Existing methods may not fully capture the complexities of facial growth patterns.

Purpose of the Study:

• To develop a novel Bayesian network approach for classifying facial biotypes.
• To classify patients into Dolichofacial, Brachyfacial, and Mesofacial categories.
• To introduce a new learning technique for tree-augmented Naive Bayes (TAN).

Main Methods:

• Utilized a Bayesian network approach for facial biotype classification.
• Developed a novel tree-augmented Naive Bayes (TAN) learning technique.
• Applied the method to classify patients into three distinct facial biotypes.

Main Results:

• The proposed Bayesian network method demonstrated superior performance over other models in accuracy, precision, recall, and kappa.
• The model exhibited the lowest dispersion, indicating high stability and robustness.
• Achieved high accuracy values compared to competitive classifiers.

Conclusions:

• The Bayesian network classifier provides a helpful decision-making tool for orthodontists.
• The method offers accurate facial biotype classification, crucial for treatment planning.
• Interactions identified within the Bayesian network have significant orthodontic interpretations.