Using PageRank to Rank Universities by Influence
Considered by many as a key invention contributing to Google’s dominance, PageRank is one of the most influential algorithms in the tech world. The intuition behind it, as we have learned in CS 2850, is simple: in a network or graph, a node is important if other important nodes point to it. Although it was originally designed to assess the value of web-pages in an automated fashion, the basic concept of PageRank turned out to be very applicable to a variety of other fields, including university ranking.
Jose Lages and his team from the University of Franche-Comte, France modified the PageRank algorithm to produce an objective ranking system for universities by mining Wikipedia pages. Objective rankings for universities are difficult to produce; geographical, cultural and historical factors are likely to influence such rankings in ways that are difficult to gauge. Similarly, universities in English-speaking countries are advantaged, since English is the primary language in the scientific community. Lages proposes a “PageRank” process for universities, in which universities mentioned in a Wikipedia article are nodes and links in Wikipedia pages pointing to a university are used to determine ranking. Although Wikipedia pages are vulnerable to abuse, the evolution of Wikipedia over time has suggested that this may become less of a concern. Lages and his team applied this algorithm to 24 different language editions of Wikipedia, composing a top 100 university ranking list for each. These lists are combined to form the global rankings.
While these rankings are unlikely to replace conventional ranking systems, they warrant merit given their objectivity. Despite de-emphasis of rankings by the universities themselves, they are known to affect decisions by funding agencies. The author notes that the French education strategy towards higher education was very likely influenced by rankings produced by notable institutions. Thus, Lages’ re-purposed PageRank process has the potential to add to a debate with real implications in the world of research and higher education.