Sentiment analysis


We have introduced and continue to develop a collection of sentiment analysis techniques to model and discover sentiment information in social media data. We have enhanced previous content-based sentiment analysis techniques with rich graph-based technologies. As an example, we model opinion summarization problem as a leader and community detection problem in network science by modeling corpus as sentence graph. We also exploit the duality between sentiment clustering and tripartite graph clustering.

ASONAM 2013; Social Com 2013; SIGMOD 2014

Graph Analytics

We have developed a set of graph analytic algorithms which serve different purposes:

  • graph reach-ability query processing
  • approximate graph pattern query processing
  • clique enumeration
  • graph clustering for k-partite graphs
  • graph summarization for k-partite graphs
  • label propagation on k-partite graphs
  • community and module detection for conteng rich networks
  • graph partitioning based on toplogy and content
  • DASFAA 2009; SIGMOD 2010; Information Systems 2011; TODS 2011; SIGKDD 2012;TKDD 2016;ISWC 2016

    Thirdslide