Ranking nodes in evolving network is of pivotal importance for many real world applications, including Web search, quantitative research evaluation, prediction of national economic development, among many others. An important class of ranking algorithms use network representations of the data to compute the inherent value, relevance, or importance of individual nodes in the system. Widely-used examples of network-based ranking algorithms include the ranking by degree (i.e., by the number of connections received by a node) and Google's PageRank. The latter is of particular interest to the physics community since it is built on statistical-physics concepts such as diffusion and random walk.
The goal of this presentation is to introduce the main elements of network-based ranking together with its applications in diverse contexts, and to discuss how my PhD thesis addressed three main research questions: How does network evolution impact the outcomes of static ranking algorithms? How to suppress ranking biases induced by network evolution? How do different ranking algorithms perform in identifying significant papers and patents in growing citation networks? The presented results will highlight the benefits from including temporal information into the ranking algorithm. I will conclude by outlining future research directions opened by the results presented in my thesis.