Premolar Ecomorphology in Anthropoid Primates: A Machine Learning Approach.

Area of Science:

• Paleontology
• Primate Ecology
• Evolutionary Biology

Background:

• Reconstructing extinct primate diets is crucial for understanding their ecology and evolution.
• Traditional proxies like molar morphology have limitations in dietary resolution.
• Premolar morphology and advanced computational methods are underexplored for dietary inference.

Purpose of the Study:

• To identify and rank dental proxies for diet in anthropoid primates using machine learning.
• To explore the potential of premolar morphology as a dietary proxy.
• To enhance dietary reconstruction methods in paleontology.

Main Methods:

• Integration of Random Forest (RF) machine learning and comparative phylogenetic methods.
• Quantification of dietary trends in premolar topography and cusp relief.
• Application of feature selection to optimize classification accuracy.

Main Results:

• Premolar protoconid relief is a strong predictor of dietary category, outperforming traditional molar and incisor proxies.
• Sexually dimorphic dietary trends were identified in honing premolars.
• Machine learning models incorporating multi-tooth data achieved the highest classification accuracy (5%-11% improvement).

Conclusions:

• Premolar morphology provides robust new dental proxies for dietary inference in extinct primates.
• Machine learning and a multi-tooth approach significantly advance ecomorphological research.
• This framework aids in clarifying ecological pressures shaping primate evolution, including the human lineage.