Sifting useful comments from Flickr Commons and YouTube

Sifting useful comments from Flickr Commons and YouTube

Abstract

Cultural institutions are increasingly contributing content to social media platforms to raise awareness and promote use of their collections. Furthermore, they are often the recipients of user comments containing information that may be incorporated in their catalog records. However, not all user-generated comments can be used for the purpose of enriching metadata records. Judging the usefulness of a large number of user comments is a labor-intensive task. Accordingly, our aim was to provide automated support for curation of potentially useful social media comments on digital objects. In this paper, the notion of usefulness is examined in the context of social media comments and compared from the perspective of both end-users and expert users. A machine-learning approach is then introduced to automatically classify comments according to their usefulness. This approach uses syntactic and semantic comment features while taking user context into consideration. We present the results of an experiment we conducted on user comments collected from Flickr Commons collections and YouTube. A study is then carried out on the correlation between the commenting culture of a platform (YouTube and Flickr) with usefulness prediction. Our findings indicate that a few relatively straightforward features can be used for inferring useful comments. However, the influence of features on usefulness classification may vary according to the commenting cultures of platforms.

Grafik Top
Authors
  • Momeni Roochi, Elaheh
  • Haslhofer, Bernhard
  • Tao, Ke
  • Houben, Geert-Jan
Grafik Top
Shortfacts
Category
Journal Paper
Divisions
Multimedia Information Systems
Subjects
Multimedia
Journal or Publication Title
International Journal on Digital Libraries
ISSN
1432-5012
Publisher
Springer Berlin Heidelberg
Date
2014
Export
Grafik Top