favorite0Second, we evaluate the dynamics of recommendation networks by simulating three different navigation models, namely Point-To-Point Search, Information Foraging and Berrypicking.
favorite5While eccentricity values are shorter for collaborative filtering (CF) than for content-based (CB) recommendation networks, distances are too long for navigation in real-world systems in both types of networks.
favorite3On our datasets, we find that recommendation networks generated by collaborative-filtering algorithms perform better for most navigation scenarios.
favorite9Navigability: We then evaluate practical navigability of recommendation networks using three different navigation models established in the literature: (i) Pointto-Point Search  as an example of goal-oriented navigation with a single fixed goal, (ii) Berrypicking  as an example of goal-oriented navigation with multiple and variable goals, and (iii) Information Foraging  as an example of exploration.
favorite4Our work extends from one-click-based evaluations of recommender systems towards multi-click analysis (i.e., sequences of dependent clicks) and presents a general, comprehensive approach to evaluating navigability of arbitrary recommendation networks..
favorite38The figures depict the ratio between stationary probabilities of pages for uniform, pragmatic and lateral random surfer.
favorite1Further, the heat maps depicted in Figure 2 strengthen our findings, as the lateral random surfer, representing users entering the website from for instance search engines, exhibits higher probabilities to visit pages which are rated as unimportant by the uniform or the pragmatic random surfer.
favorite2In particular, we are interested in analyzing how real users assess the importance of Web pages for navigation and how that assessment compares to that of the random surfer.
favorite2Our high level contributions are (i) a comparison of stationary distributions of different types of the random surfer to quantify the similarities and differences between those models as well as (ii) new insights into the impact of search engines on traditional user navigation.
favorite9ABSTRACT The random surfer model is a frequently used model for simulating user navigation behavior on the Web. Various algorithms, such as PageRank, are based on the assumption that the model represents a good approximation of users browsing a website.