Authors: Stephanie Clark*,
Topics: Spatial Analysis & Modeling, Quantitative Methods
Keywords: Machine Learning, Big Data, MaxEnt, Maximum Entropy, Predictive Model, Archaeology, Analytical model, GeoSemantics, ESRI ArcPro, Natural Language Processing (NLP), Neural Network, Data Mining, Text Mining,Raster Imagery
Session Type: Poster
Presentation File: No File Uploaded
Archaeological predictive modeling is an essential tool for cultural resource management and historic preservation agencies throughout the state of Texas. By taking a big data approach and combining several machine learning-based algorithms, the proposed study will allow Texas’ archaeological data to ‘speak for itself.’ Natural language processing (NLP) will classify syntax from recorded archaeological site information to derive real-world geographical and contextual meaning. The unconstrained environmental descriptions will overlay several raster layers using ESRI’s ArcPRO in conjunction with Maximum Entropy (MaxEnt) software to verify the model’s accuracy. The output from this process serves as a semi-supervised training dataset to ultimately use within an unsupervised classification based on all unknown areas within Texas. When we let the story of Texas’ past unfold through the geo-narrative of the archaeological record, the resulting model is heavily rooted in theory-based decisions. As such, this model will reaffirm a combined method approach by highlighting the potential for a truly unique archaeological interpretation of Texas.
To access contact information login