Digging Deeper with Data: Archaeological Predictive Modeling Using Novel Techniques in Artificial Intelligence and Machine Learning

Abstract

Here, we aim to explore archaeological information with a data science approach. Specifically, we intend to use cutting-edge predictive modeling techniques, where the primary objective of these techniques is to harness mathematical tools and machine learning, mainly using TensorFlow neural networks, to predict and anticipate potential locations of yet-to-be-discovered archaeological sites. We will consider factors from existing site data and environmental indicators, including proximity to water sources, elevation, and soil types.

We will analyze Iron Age Irish ringforts—circular fortified settlements found primarily in Ireland. These sites served as residences, defensive structures, and indicators of social status during the Early Medieval period. We are keen to determine if contemporary AI techniques can outpace traditional methodologies in this field. Expected outcomes include hosting tailored workshops, publishing our findings, and developing tools to assist local archaeologists and, potentially, policymakers with these methodologies. The proposed research promises deeper insights into ancient civilizations and bolsters NMU’s Data Science and GIS programs, underlining NMU’s commitment to groundbreaking, interdisciplinary research. With its rich archaeological data records, Ireland is the ideal training ground for these innovative techniques. Furthermore, our methodologies may be adjusted for North American contexts, particularly in regions where archaeological records remain limited and fragmented.