Read all about it! : a research support tool for automatically finding relevant prior and related work and sequencing it for reading

Abstract

This dissertation describes a research tool that automatically finds prior work associated with a research paper. A paper is inputted into the tool and by iterating through the references, the tool builds a dependency graph of papers and determines a progression in which they should be read. This ensures that subjects introduced in later papers but explained in earlier papers are understood by the user. The PageRank algorithm, along with the Katz Distance, vector-space models and other algorithms will rank each papers’ importance. Less important papers can be omitted from the timeline and a visualization will indicate to the user which papers are more important than others. Since no Gold Standard is available, this project is a proof of concept. However, an evaluation was conducted using a Silver Standard, where multiple dependency graphs of virtual papers were used as testing mediums for this project. The aim of this project is to make studying easier and more fun for academics. The research tool uses Google Scholar as its primary digital library, therefore, it has the potential to help many academics studying a wide variety of subjects.