Predicting time-to-event from Twitter messages

Hannah Tops, Antal Van Den Bosch, Florian Kunneman

Research output: Chapter in Book / Report / Conference proceedingConference contributionAcademicpeer-review

Abstract

We describe a system that estimates when an event is going to happen from a stream of microtexts on Twitter referring to that event. Using a Twitter archive and 60 known football events, we train machine learning classifiers to map unseen tweets onto discrete time segments. The time period before the event is automatically segmented; the accuracy with which tweets can be classified into these segments determines the error (RMSE) of the time-to-event prediction. In a cross-validation experiment we observe that support vector machines with χ2 feature selection attain the lowest prediction error of 52.3 hours off. In a comparison with human subjects, humans produce a larger error, but recognize more tweets as posted before the event; the machine-learning approach more often misclassifies a 'before' tweet as posted during or after the event.

Original languageEnglish
Title of host publicationBNAIC 2013
Subtitle of host publicationProceedings of the 25th Benelux Conference on Artificial Intelligence
Pages207-214
Number of pages8
Publication statusPublished - 1 Jan 2013
Externally publishedYes
Event25th Benelux Conference on Artificial Intelligence, BNAIC 2013 - Delft, Netherlands
Duration: 7 Nov 20138 Nov 2013

Publication series

NameBelgian/Netherlands Artificial Intelligence Conference
ISSN (Print)1568-7805

Conference

Conference25th Benelux Conference on Artificial Intelligence, BNAIC 2013
CountryNetherlands
CityDelft
Period7/11/138/11/13

Fingerprint

Dive into the research topics of 'Predicting time-to-event from Twitter messages'. Together they form a unique fingerprint.

Cite this