Unsupervised discovery of non-trivial similarities between online communities
Abstract: Highlights•We introduce a new algorithm to quantify the similarity between online communities.•We compare the results of the new algorithm to three competitive methods.•The new algorithm better discovers non-trivial similarities between communities.•We present a new human-annotated dataset of similarities between Reddit communities.
Loading