Abstract by Tyler Jones
Random walks, page rank, and non-backtracking random walks
Random walks through a network are often used in algorithms to estimate node importance, the most well known of these examples is the page rank metric. Page rank gives a score to each graph vertex based on how many incoming links it has. We explore how page rank values change when we constrain our random walks to be non-backtracking, and what implications this has on their ranking and other spectral properties of the network.