Comparison of data mining algorithms for sex determination based on mastoid process measurements using cone-beam...

Area of Science:

• Forensic Anthropology
• Radiology
• Data Science

Background:

• The mastoid region's anatomical position makes it suitable for studying sexual dimorphism.
• Accurate sex determination is crucial in forensic investigations.

Purpose of the Study:

• To determine sex in the Iranian population using mastoid process measurements.
• To evaluate the effectiveness of various data mining algorithms for sex determination.

Main Methods:

• Retrospective analysis of 190 3D cone-beam computed tomographic (CBCT) images.
• Measurement of 9 specific landmarks on the mastoid process.
• Construction and cross-validation of multiple predictive models using data mining algorithms.

Main Results:

• Statistically significant differences between sexes were found in most mastoid variables.
• The random forest model demonstrated the highest accuracy (97.0%) in sex prediction.
• Intermastoid distance (IMD) and the distance between the most prominent convex mastoid point (IMSLD) were key predictors.

Conclusions:

• Data mining algorithms can be effectively used for sex determination based on mastoid process morphology.
• This approach offers a potential tool for forensic applications.
• Further development could lead to accurate and reliable forensic sex determination methods.