Mandibular and dental measurements for sex determination using machine learning

Area of Science:

• Forensic Anthropology
• Biometrics
• Machine Learning Applications

Background:

• Sex determination is crucial in forensic anthropology.
• Traditional methods rely on skeletal and dental analysis.
• Integrating morphometric data with advanced algorithms offers new possibilities.

Purpose of the Study:

• To evaluate the efficacy of combining mandibular and dental dimensions for sex determination.
• To assess the performance of various machine learning algorithms in this task.
• To identify key measurements with high predictive power.

Main Methods:

• Utilized lateral cephalograms and dental casts from 108 individuals.
• Extracted three mandibular and eight dental mesio-distal dimensions.
• Applied univariate statistics for variable selection and trained seven machine learning models.
• Validated models using threefold cross-validation and ROC curve analysis.

Main Results:

• Mandibular ramus height and lower first molar mesio-distal size showed significant predictive capability.
• Model accuracy ranged from 0.58 to 0.79 in testing.
• Logistic regression achieved the highest performance with an Area Under the Curve (AUC) of 0.84.

Conclusions:

• The combination of mandibular and dental dimensions is a promising approach for sex prediction.
• Machine learning techniques demonstrate significant potential as tools in forensic sex determination.
• Further research is warranted to validate these findings and explore clinical applications.