Principal Researcher: Elisavet Charalambous (G. M. EUROCY Innovations LTD, Cyprus)
LinkedIn profile: cy.linkedin.com/pub/liza-charalambous/24/6a8/72/
This project is concerned in field segments involving the development and application of Computational Intelligence and other advanced statistical learning techniques along with pattern recognition techniques in reinforcing analysis by identifying internal structures in archaeological data. Products of this work are envision engagement by the scientific and archaeological community.
The Current State:
Technology serves archaeology in copious means including the development of instruments and methods assisting in the unearthing of findings, their subsequent analysis as well as the interpretation of produced results through behavioral studies. This is usually done through the means of compositional and other analysis followed by interpretation of the results.
Statistical Analysis Challenges:
Over the last few decades, the deployment of advanced statistical methods has been proposed in assisting answering various archaeological questions. Questions such the categorization of artefacts based on their origin, technology, dating etc. are common and very likely to be answered with the deployment of statistical learning methods.
Despite significant progress, key questions including spatiotemporal analysis, utilization of sparse samples as well as the modeling of exhibited uncertainties have not been adequately resolved.
The development of techniques that assist the analysis of heterogeneous data with the aim to utilize inherent information which cannot be considered with the use of a standard analysis method. The highly interdisciplinary nature of this challenge combines fundamental practices from (but not limited to) the fields of archaeology, statistics and technology (computer science and artificial intelligence, computer engineering and computational intelligence).
Findings can, nowadays, be recorded and examined from different perspectives and detail levels with the utilization of various technological methods. Since both structured and semi-structured data are recorded, important inherent information can be exploited which upon appropriate analysis and interpretation assists providing knowledge to experts; a central motivator of this project.
The new era develops systems which add layers of abstraction in managing stored information; data can now be accessed through a number of different application/ services. Additionally the deployment of intelligent techniques such as neural network structures (MPL, SOM, LVQ) and Fuzzy Logic allow the discrimination of highly complex and non-linearly separable regions.
Open Source Toolkit:
Part of this project is concerned with the implementation of an archaeological toolkit implemented in MATLAB, a well-supported and widely accessible mathematical software package.
Research results will be accessible through this solution which will allow the user to deploy with ease several statistical learning and CI methods for data analysis. The developed algorithms are implemented based on each method’s underlying mathematical structure and allow parameter flexibility where needed with main consideration to increase the user’s engagement.