Female heads, ivory. Nimrud, c. 900—700 BCE. The British Museum, WA 118200, WAS 118232, WA 118233, © Trustees of The British Museum
This code identifies stylistically similar groups in a corpus of Levantine ivory statuettes using maximum likelihood and maximum posterior clustering. While it has been written with the application to art historical data in mind, it may be used generally to perform clustering data with a combination of categorical and real-valued features, where some data may be missing.
Data may be read from a CSV file. Each row is interpreted as a separate observation. Numerical columns are assumed to be real-valued, string-valued columns are assumed to be categorical.
The repository aneic-core contains the sources for the statistical algorithm, as well as the levantine corpus dataset.
The repository aneic-scripts contains support code needed in order to perform analysis of the Levantine corpus.
git clone https://github.com/aneic/aneic-scripts.git
cd aneic-scripts/130317_01_mfm_sigma_1.0_a_2.0
git clone https://github.com/aneic/aneic-core.git
python run_mfm.py
This project is both open data and open source.
The levantine corpus dataset is copyright 2013 Amy Gansell (@amygansell). It is distributed under the Creative Commons Attribution-ShareAlike 3.0 Unported License.
The source code is copyright 2013 Jan-Willem van de Meent (@jwvdm) and Sakellarios Zairis (@szairis). It is distributed under MIT license.
Please cite this work as:
Amy R. Gansell, Jan Willem van de Meent, Sakellarios Zairis, and Chris H. Wiggins, "Stylistic clusters and the Syrian/South Syrian tradition of first millennium BCE Levantine ivory carving: a machine learning approach," Journal of Archaeological Science 44 (2014): 194-205.