He, Yulan; Lin, Chenghua and Cano Basave, Amparo
Due to copyright restrictions, this file is not available for public download
Click here to request a copy from the OU Author.
|Google Scholar:||Look up in Google Scholar|
We propose a dynamic joint sentiment-topic model (dJST) which is able to effectively track sentiment and topic dynamics over the streaming data. Both topic and sentiment dynamics are captured by assuming that the current sentiment topic speciﬁc word distributions are generated according to the word distributions at previous epochs. We study three different ways of accounting for such dependency information, (1) Sliding window where the current sentiment-topic-word distributions are dependent on the previous sentiment topic speciﬁc word distributions in the last S epochs; (2) Skip model where history sentiment topic-word distributions are considered by skipping some epochs in between; and (3) Multiscale model where previous long- and short- timescale distributions are taken into consideration. We derive efﬁcient online inference procedures to sequentially update the model with newly arrived data and show the effectiveness of our proposed model on the Mozilla add-on reviews crawled between 2007 and 2011.
|Item Type:||Conference Item|
|Copyright Holders:||2012 IEEE|
|Academic Unit/Department:||Knowledge Media Institute|
|Interdisciplinary Research Centre:||Centre for Research in Computing (CRC)|
|Depositing User:||Kay Dave|
|Date Deposited:||30 Oct 2012 11:22|
|Last Modified:||30 Oct 2012 22:00|
Actions (login may be required)
|Public: Report issue / request change|