1 / 8

Towards the Question of Redundancy in Social Tagging Networks

Towards the Question of Redundancy in Social Tagging Networks. Social Tagging. Users annotate things (resources) with labels (tags) These annotations are shared, creating a collaborative dataset called a folksonomy. coffee. User A. programming. User B. java.

min
Download Presentation

Towards the Question of Redundancy in Social Tagging Networks

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Towards the Question of Redundancy in Social Tagging Networks

  2. Social Tagging • Users annotate things (resources) with labels (tags) • These annotations are shared, creating a collaborative dataset called a folksonomy coffee User A programming User B java

  3. Recommendation in Folksonomies • With some knowledge about a tagging user, we want to present that user with new information: • Given a resource, what are some tags that the user may find appropriate for annotation? • Given a tag (or set of tags), what are some relevant resources the user may find interesting? • Generally, what resources may interest the user? • What other users have similar interests to this user? • The first of these has been the focus of most recommender research.

  4. Tag Recommendation • Task: Recommend tags during the annotation process • Benefits • Reduce user effort • Encourage tags more frequently • Encourage the application of more tags • Reuse common tags • Offer tags the user had not previously considered • Reduce noise created by capitalization inconsistencies, spelling errors and other discrepancies. • Results: A cleaner denser dataset that is useful in its own right or for future data mining techniques.

  5. Data Structure UR UT TR

  6. Techniques for Tag Recommendation • Memory-based • Simple Popularity contest • K-Nearest Neighbor, adapted from CF • Model-based • Link Analysis (Adapted PageRank, FolkRank) • Latent Concept Models • LSA (using SVD or Hebbian deflation) • pLSA (using learned belief networks • Knowledge-based • Use external ontology to relate tags/resources

  7. Evaluation • What makes a “good” tag recommendation? • According to IR tradition, we measure the predictive power of a recommendation algorithm

  8. The Real Motivation • Consider the following recommendations for a resource representing a research paper on folksonomies • Is prediction really what we want in a practical recommendation system?

More Related