Development of a Digital Comparative Collection of Chert Types in Ontario and the Evaluation of Change in Accuracy and Confidence of Chert Type Identifications

Description

The objective of this thesis is to create a foundation for a digital comparative collection of chert types found on archaeological sites in Ontario, both local and non-local varieties, and to evaluate the impact of a digital reference collection on the confidence and accuracy of the user in comparison to hard copy guides or hand samples that are more often traditionally used. Spatial and temporal variation in the use of different lithic raw materials has shown to provide insight into cultural interaction, resource exchange and control across multiple periods in the study of Indigenous archaeology; however comparative collections needed to conduct analyses remain accessible only in a physical form. This study will build a foundation, develop a prototype using a represented sample of hand specimens from the William Fox Northeastern North American Lithic Reference Collection (referred to hereafter as The Fox Collection) at Trent University, and create a prototype digital system to assist the user in identifying the chert type through the use of a simple expert system using a decision tree. The digital identification system was tested by a group of volunteers with to compare accuracy and confidence in analysis against traditional methods of hand samples and hard copy guides. When supplied with the digital reference collection, a statistically significant improvement in the accuracy and confidence of chert identification was identified.
Author Keywords: database design, digital comparative collection, digital identification system, expert system, Ontario archaeology, raw material analysis