Predicting the popularity of online articles based on user comments

Abstract : Understanding user participation is fundamental in anticipating the popularity of online content. In this paper, we explore how the number of users' comments during a short observation period after publication can be used to predict the expected popularity of articles published by a countrywide online newspaper. We evaluate a simple linear prediction model on a real dataset of hundreds of thousands of articles and several millions of comments collected over a period of four years. Analyzing the accuracy of our proposed model for different values of its basic parameters we provide valuable insights on the potentials and limitations for predicting content popularity based on early user activity.
Document type :
Conference papers
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-00940014
Contributor : Fabien Mathieu <>
Submitted on : Friday, January 31, 2014 - 11:11:35 AM
Last modification on : Thursday, March 21, 2019 - 1:00:13 PM

Identifiers

Citation

Alexandru Tatar, Jeremie Leguay, Panayotis Antoniadis, Arnaud Limbourg, Marcelo Dias de Amorim, et al.. Predicting the popularity of online articles based on user comments. WIMS 2011 - 1st International Conference on Web Intelligence, Mining and Semantics, May 2011, Sogndal, Norway. pp.67:1--67:8, ⟨10.1145/1988688.1988766⟩. ⟨hal-00940014⟩

Share

Metrics

Record views

317