Understanding retweeting behaviors in social networks
CIKM '10 Proceedings of the 19th ACM international conference on Information and knowledge management
Proceedings of the 20th international conference on World wide web
Examining Multi-factor Interactions in Microblogging Based on Log-linear Modeling
ASONAM '12 Proceedings of the 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012)
Hi-index | 0.00 |
In this paper, we consider a fundamental social network issue that illustrates how information dynamically flows through a social media network. Inferring the number of times that a particular message posted by some specific user will be retweeted by his followers and predicting the number of readings of the posted message via various retweeting chains are central to understanding the underlying mechanism of the retweeting behaviors. Specifically we work on the Task 2 of the WISE 2012 Challenge, i.e., predicting retweet behaviors in the Sina Webo data set. We develop an approach based on the Autoregressive-Moving-Average (ARMA). In the approach, we treat retweeting activities of each original tweet as a time series where each value corresponds to the number of times that the original tweet is tweeted or the number of times of possible-view of the original tweet during that particular time period. For each tweet in the test data, our approach first identifies the most similar message from the training data based on the similarity between their time series values in the same length period as provided in the test tweet, fits the ARMA models over the whole time series of the identified message, and then applies the fitted model over the time series of the test tweet to predict future values. We report our prediction results and findings in this paper.