FolkTrails: Interpreting Navigation Behavior in a Social Tagging System

Thomas Niebler, Martin Becker, Daniel Zoller, Stephan Doerfel, & Andreas Hotho (CIKM-2016)
In the paper “FolkTrails: Interpreting Navigation Behavior in a Social Tagging System”, Niebler et al. present an investigate navigation trails in the popular scholarly social tagging system BibSonomy. They tried to understand users’ dynamic browsing behaviors on this kind of systems, how individual users behave differently on navigation, and whether the semantic nature of the underlying folksonomy can help to explain navigation.
Firstly, Niebler et al. study different hypotheses about the navigational user behavior in Bibsonomy, a social bookmarking system. They test these hypotheses by using HypTrails, an approach utilizing first-order Markov chain models and Bayesian inference for expressing and comparing hypotheses about human trails. These hypotheses include:
  • Uniform hypothesis: serves as a baseline, users randomly pick a link to navigate. Any “real” hypothesis capturing a structurally interesting aspect of user behavior, will exhibit a higher evidence than this simple hypothesis.
  • Page consistent hypothesis: users make a transition from a page to itself.
  • Category consistent hypothesis: transitions between two pages often occur between pages of the same category.
  • User consistent hypothesis: a transition’s target and source page belong to the same user.
  • Folksonomy consistent hypothesis: Social tagging systems map links of the underlying folksonomy to actual hyperlinks of the system.
  • Semantic navigation hypothesis: they compute the similarities between two pages by using the tag clouds belonging to the pages to measure cosine similarity.
Moreover, to analyze the mutual relationships between these hypotheses, they also evaluate combined hypotheses which include:
  • Folksonomy consistent & semantic navigation hypothesis
  • User consistent & semantic navigation hypothesis
  • User consistent & folksonomy navigation hypothesis
Niebler et al. collect data from Bibsonomy system. After preprocessing and filtering, they have 456,777 bookmark posts and 2,410,844 publication posts using 65,228 distinct tags. The “Request Log” Dataset about user transition consists of 103,415 distinct visited pages, 327,060 transitions between these pages, 123,452 transitions were self-transitions and 261,300 were own-transitions. In order to evaluate Page Consistent hypothesis, they also categorize six page types which are grouped into three categories corresponding to tag, resource, and user.
In the evaluation part, they firstly compare all hypotheses with overall Request Log dataset. Then, they separate the dataset into subsets to see whether not different users or different kinds of navigation perform differently. They divide the dataset by considering inside and outside navigation; users’ gender; short-term and long-term users; and tagger classes (i.e., categorizers-describers, generalists-specialists). Based on the results, Niebler et al. claim that:
  • Users mainly navigate on their own resources,
  • The semantic component influences navigation behavior on BibSonomy,
  • Users fall back to the folksonomy structure when browsing outside of their own pages,
  • Different genders did not exhibit interesting behavioral deviations, and
  • Short-term users and different tagging types follow certain behavioral patterns matching their individual characteristics.
With the emerging of social information access, many studies focus on different aspects of social tagging systems. Navigation is one of the most important ones on this kind of systems. This paper provides us of a better understanding of users’ navigation behavior on a social tagging system so that we are able to assess the effectiveness of a navigational concept and improve it to better meet the users’ information needs.


Popular posts from this blog

[Talk Summary] Machine Learning and Privacy: Friends or Foes?

SoRec: Social Recommendation Using Probabilistic Matrix Factorization

[Talk Summary 3] Personalized Recommendations using Knowledge Graphs