Your new post is loading...
From the original research paper by Derek Greene, Gavin Sheridan, Barry Smyth, and Padraig Cunningham of the School of Computer Science & Informatics, University College Dublin entitled "Aggregating Content and Network Information
to Curate Twitter User Lists": "The problem of content curation in social media networks is becoming increasingly important, particularly in the context of news curation for media outlets.
In the case of Twitter, curating a list of authoritative users tweeting about a
given news story provides a means of monitoring discussions around that story.
However, currently this is a time-consuming manual task. Here we presented a
range of criteria for building topical user lists, based on an initial seed set.
By analysing the cohesion of the training data across di erent views of the same
datasets, we demonstrated the strengths and weaknesses of these recommenda-
tion criteria, in the context of a limited availability of Twitter data.
To overcome the weaknesses, we proposed the use of SVD rank aggregation. Experiments on a range of Twitter datasets relating to US politics demonstrated that this aggregation process yields more robust recommendations, succeeding in cases where individual content- or network-based criteria perform poorly."
http://arxiv.org/pdf/1206.1728.pdf
Via Paulo Simões
