When Google's algorithm fails: Bringing order to social networks

This paper presents a novel social network ranking algorithm called CrRank, designed specifically for social media. The algorithm computes the user ranks by analyzing the content they contribute in terms of words and hashtags. The main intuition is based on the user subscriptions to the various topics of a given dataset. Such subscription is measured against a notion of ground-truth that is objectively computed in advance. CrRank is a biologically-inspired multilayered network algorithm. It is equipped with a network measure (i.e., density) and social media heuristics (e.g., number of hashtags votes) to compute distinct ranks. The algorithm and experiments have shown a great deal of promise in outperforming major ranking algorithms (e.g., PageRank) and other measures (e.g., network centrality).