Skip to main content
Fig. 3 | Computational Social Networks

Fig. 3

From: A method for evaluating discoverability and navigability of recommendation algorithms

Fig. 3

Bow tie membership over N. The figures depict the memberships of items to the components of the bow tie model (see Fig. 2), and the changes to membership as the number of recommendations N at each item increases from one through twenty. Each component is of the size proportional to the number of items it contains, with all components adding up to \(100\%\). Except for the association rules recommendations, most nodes belong to the IN or SCC component. This is beneficial for discoverability, as this means there exists paths between most pairs of items (in case of SCC), or that, at least, almost any item in the system allows reaching the strongly connected set of items

Back to article page