Introducing Paleoethnobotany to Machine Learning: A Case Study in the Genus *Capsicum

Summary

This is an abstract from the "SAA 2021: General Sessions" session, at the 86th annual meeting of the Society for American Archaeology.

Chili peppers (*Capsicum spp.) are an incredibly diverse and abundant crop across the Americas whose domestication began around 10,000 BP as a complex co-evolutionary process between humans and these plants. This genus has served many culinary, medicinal, and ritualistic uses throughout its evolution and diversification. With an interest in tracking the domestication of the *Capsicum genus over time, we seek to develop a method of species-level identification based on seed morphometrics. To this end, we test a variety of machine learning algorithms on the collected morphometric data to determine which best models the high-dimensional space of the data. Machine learning algorithms utilized are linear discriminant analysis, neural networks, decision tree, Random Forest, as well as a series of dimensionality reduction algorithms to prevent overfitting. We also compare these models to previous models developed for species-level identification of *Capsicum seeds. As the only remaining portions of chili pepper at archaeological sites are very often the seeds, these species-level identification models can be utilized in the field to identify the species of seeds found in order to track their domestication over time.

This Resource is Part of the Following Collections

• Society for American Archaeology 86th Annual Meeting, Online (2021)   •
• SAA 2021: General Sessions

Keywords

Spatial Coverage

Record Identifiers

Abstract Id(s): 32472