Theses and Dissertations Collection

Information is often tagged with metadata that indicates access guidance and reliability criteria to protect data from unauthorized access or release. In the US government, this often takes form as a security classification. New documents are classified by an original classifier, and subsequent documents derived from the original are classified using a derivative classification process. Derivative classification is a process where a derivative classifier applies a security classification guide, generated by an original classifier, to a new document to apply the correct classification levels to this document. This is often a time consuming and tedious process.

Modern computer systems have the ability to sift through large amounts of data, dynamically generating content based on attributes of the user, including current search terms, history, location and related information. This is all derived information that needs to be correctly classified. In addition, conflicting guidelines can be present in security classification guides, and mistakes can occur in the process. Therefore there is a need for a process to automate the derivative classification process, eliminate errors and simplify the process.

The goal of this thesis is to provide a solution to the derivative classification problem by creating an automated derivative classification process and associated tool. This is demonstrated through development of a mechanism for original classifiers to create a rule file that can be used to automate the derivative classification process and is then incorporated into the UITags project, a broad research project examining different security metadata technologies by inserting metatags in XML documents.