Document Triage for Semi-Structured Documents

The generic term for looking for information is labelled Information Seeking. The process entails everything from the realisation that a gap in our knowledge exists, to the satisfying or abandoning of that need. This process has been researched thoroughly and models representing the conceptualisation of the process are in abundance.

This project focuses on a sub-process of information seeking. The process is called “Document Triage”. Simply speaking, document triage is the process we go through to make relevance decisions on documents. It occurs after a document set is presented to us, for example after a query is executed on a search engine, and ends when the documents are either all rejected or when relevant information is found. Anything from our physical actions to our cognitive workings are used to provide our own individual triage activities.

Document triage takes place in physical as well as electronic environments. An information seeker can search for academic documents using a search engine, to be returned with suitable results in a list. Similarly, an information seeker can request for academic documents by using a physical library. The librarian can then present the individual with an appropriate shelf or section of the library, which might prove useful to the information need. Both of these scenarios have one thing in common; namely, that the information seeker is then liable to triage through the information in order to extract what he or she thinks is relevant.

Our focus in this project is professional search on semi-structured documents. This includes documents from fields such as academia, law (case law) and patents.

The main objectives are:

a) investigate the visual attention of information seekers performing document triage on semi-structured documents using electronic means

b) produce conceptual models and theoretical frameworks from empirical and qualitative data

This project began as part of my PhD work with Dr. George Buchanan.