Recruiting a skeleton crew-Methods for simulating and augmenting paleoanthropological data using Monte Carlo based...

Area of Science:

• Paleoanthropology
• Evolutionary Biology
• Computational Biology

Background:

• Data scarcity and quality are significant challenges in human evolutionary studies.
• Limited fossil data impedes crucial analyses like classification and predictive modeling.

Purpose of the Study:

• To present Monte Carlo methods for simulating paleoanthropological data.
• To demonstrate the enhancement of datasets using realistic synthetic data.
• To introduce an R library, AugmentationMC, for data simulation.

Main Methods:

• Utilized Monte Carlo based methods, including Markov Chain Monte Carlo, for data simulation.
• Applied methods to cross-sectional biomechanical and geometric morphometric 3D landmark datasets.
• Emphasized Machine Teaching over Machine Learning for 3D model simulation.

Main Results:

• Generated statistically equivalent synthetic morphometric data using Monte Carlo algorithms.
• Demonstrated that simulated data enhances datasets for complex tasks like classification.
• Showcased Monte Carlo methods outperforming bootstrapping for data simulation.

Conclusions:

• Monte Carlo based simulation offers a powerful tool for handling paleoanthropological data limitations.
• Synthetic datasets can augment real data, advancing research in human evolution.
• The AugmentationMC library provides accessible algorithms for paleoanthropological data simulation.