Classifying Short Messages on Social Networks using Vector Space Models

Ricardo Lage, Peter Dolog, Martin Leginus

2013

Abstract

In this paper we propose a method to classify irrelevant messages and filter them out before they are published on a social network. Previous works tended to focus on the consumer of information, whereas the publisher of a message has the challenge of addressing all of his or her followers or subscribers at once. In our method, a supervised learning task, we propose vector space models to train a classifier with labeled messages from a user account. We test the precision and accuracy of the classifier on over 13,000 Twitter accounts. Results show the feasibility of our approach on most types of active accounts on this social network.

Paper Citation

in Harvard Style

Lage R., Dolog P. and Leginus M. (2013). Classifying Short Messages on Social Networks using Vector Space Models . In Proceedings of the 9th International Conference on Web Information Systems and Technologies - Volume 1: WEBIST, ISBN 978-989-8565-54-9, pages 413-422. DOI: 10.5220/0004357304130422

in Bibtex Style

@conference{webist13,
author={Ricardo Lage and Peter Dolog and Martin Leginus},
title={Classifying Short Messages on Social Networks using Vector Space Models},
booktitle={Proceedings of the 9th International Conference on Web Information Systems and Technologies - Volume 1: WEBIST,},
year={2013},
pages={413-422},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0004357304130422},
isbn={978-989-8565-54-9},
}

in EndNote Style

TY - CONF
JO - Proceedings of the 9th International Conference on Web Information Systems and Technologies - Volume 1: WEBIST,
TI - Classifying Short Messages on Social Networks using Vector Space Models
SN - 978-989-8565-54-9
AU - Lage R.
AU - Dolog P.
AU - Leginus M.
PY - 2013
SP - 413
EP - 422
DO - 10.5220/0004357304130422