TY - GEN
T1 - A network-based model for predicting hashtag breakouts in twitter
AU - Alzahrani, Sultan
AU - Alashri, Saud
AU - Koppela, Anvesh Reddy
AU - Davulcu, Hasan
AU - Toroslu, Ismail
N1 - Publisher Copyright:
© Springer International Publishing Switzerland 2015.
PY - 2015
Y1 - 2015
N2 - Online information propagates differently on the web, some of which can be viral. In this paper, first we introduce a simple standard deviation sigma levels based Tweet volume breakout definition, then we proceed to determine patterns of re-tweet network measures to predict whether a hashtag volume will breakout or not. We also developed a visualization tool to help trace the evolution of hashtag volumes, their underlying networks and both local and global network measures. We trained a random forest tree classifier to identify effective network measures for predicting hashtag volume breakouts. Our experiments showed that “local” network features, based on a fixed-sized sliding window, have an overall predictive accuracy of 76%, where as, when we incorporate “global” features that utilize all interactions up to the current period, then the overall predictive accuracy of a sliding window based breakout predictor jumps to 83%.
AB - Online information propagates differently on the web, some of which can be viral. In this paper, first we introduce a simple standard deviation sigma levels based Tweet volume breakout definition, then we proceed to determine patterns of re-tweet network measures to predict whether a hashtag volume will breakout or not. We also developed a visualization tool to help trace the evolution of hashtag volumes, their underlying networks and both local and global network measures. We trained a random forest tree classifier to identify effective network measures for predicting hashtag volume breakouts. Our experiments showed that “local” network features, based on a fixed-sized sliding window, have an overall predictive accuracy of 76%, where as, when we incorporate “global” features that utilize all interactions up to the current period, then the overall predictive accuracy of a sliding window based breakout predictor jumps to 83%.
KW - Diffusion networks
KW - Hashtag volumes
KW - Information diffusion
KW - Prediction
KW - Social networks
UR - http://www.scopus.com/inward/record.url?scp=84925303347&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84925303347&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-16268-3_1
DO - 10.1007/978-3-319-16268-3_1
M3 - Conference contribution
AN - SCOPUS:84925303347
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 3
EP - 12
BT - Social Computing, Behavioral-Cultural Modeling, and Prediction - 8th International Conference, SBP 2015, Proceedings
A2 - Xu, Kevin
A2 - Agarwal, Nitin
A2 - Osgood, Nathaniel
PB - Springer Verlag
T2 - 8th International Conference on Social Computing, Behavioral-Cultural Modeling, and Prediction, SBP 2015
Y2 - 31 March 2015 through 3 April 2015
ER -