Predicting discussions on the social semantic web

From Brede Wiki
Jump to: navigation, search
Conference paper (help)
Predicting discussions on the social semantic web
Authors: Matthew Rowe, Sofia Angeletou, Harith Alani
Citation: The Semantic Web: Research and Applications 6644 in Lecture Notes in Computer Science : 2011 June
Editors: Grigoris Antoniou, Marko Grobelnik, Elena Simperl, Bijan Parsia, Dimitris Plexousakis, Pieter de Leenheer, Jeff Pan
Publisher: Springer-Verlag, Berlin Heidelberg 2011
Meeting: 8th Extended Semantic Web Conference
DOI: Define doi.
Web: DuckDuckGo Bing Google Yahoo!Google PDF
Article: Google Scholar PubMed
Restricted: DTU Digital Library
Format: BibTeX

Predicting discussions on the social semantic web describes a study on identifying key indicators for postings on Twitter that generate discussions: "Will a given post start a discussion?"

[edit] Method

Description of features connected:

These include user features: in degree, out degree, list degree, post count, user age, post rate as well content features: post length, complexity, uppercase count, readability, verb count, noubn count, adective count, referral count, time in the day, informativeness (terminological novelty), politary (using SentiWordnet)

Four machine learning methods was used: Perceptron, support vector machine, naïve Bayes classifier and J48.

[edit] Results

User list degree, user in degree, content- time in day were important featuree in discussion posts.

[edit] Related papers

  1. Anticipating discussion activity on community forums
  2. Good friends, bad news - affect and virality in Twitter
Personal tools