Text Mining and Social Media Analytics Lab
- Large-scale social media-based human behavior modeling: applying text mining and machine learning to a large sample of social media posts to gain deep insight into human behavior. Specially, we are interested in
- building computational models of personal traits (e.g., personality, needs, values, emotions and interests) and human behavior (e.g., brand preference, purchase decision and drug use)
- automated social media-based large-scale personal persuasion: adapting the content and linguistic style of a campaign message to enhance its appeal to different individuals (e.g., to motivate healthy and drug-free lifestyle) on social media.
- Text mining with humans in the loop: developing new active and interactive text mining techniques to support human-in-the-loop text mining. Specifically, we are interested in
- constrained topic modeling to allow topic models to incorporate user feedback and domain knowledge.
- active topic modeling to improve the effectiveness of eliciting human input in semi-supervised topic modeling.
- interactive text visualization to seamlessly integrate text mining with information visualization to help users explore large text corpora.
- the usability of text mining algorithms (e.g., improving the stability and explainability of topic models).
- Intelligent Interactive systems which include
- Multimedia/multimodal Interactive Systems
- Adaptive Conversation Systems
- Dr. Pan gave a talk at National Institute on Drug Abuse (NIDA) on “Social Media-based User Behavior and Trait Analysis for Drug Abuse Research” (September. 2016)
- Our paper on “Personalized Emphasis Framing for Persuasive Message Generation” was accepted by EMNLP 2016 (August, 2016)
- Dr. Pan gave a talk at Bloomberg LP Data Science Seminar (July 2016)
- Dr. Pan gave a talk at the CS department of Columbia University (July 2016)
- Our paper was presented at the IJCAI Sister Conference Best Paper Track (July 2016)
- Dr. Pan presented an invited paper at AAAI 2016 on “What’s Hot in Intelligent User Interfaces” (Feb 2016).
- Our paper on “Using Personal Traits for Brand Preference Prediction” was accepted by EMNLP 2015 (August 2015)
- Our paper on “An Uncertainty-Aware Approach for Exploratory Microblog Retrieval” was accepted by IEEE VAST 2015 (TVCG track) (July 2015)
- Our paper on “Expertise Signals for Public and Enterprise Social Q&A” was accepted by ICWSM 2015 (March 2015)
- Our paper on “User-Directed Non-Disruptive Topic Model Update for Effective Exploration of Dynamic Content” won “best paper honorable mention” at IUI 2015! (March 2015)
- Dr. Pan gave a keynote at the 4th Interactive Visual Text Analytics workshop in Atlanta (March 2015).
- Our paper “User-Directed Non-Disruptive Topic Model Update for Effective Exploration of Dynamic Content” was accepted by IUI’2015.
- Phd student Tao Ding received UMBC GSA student research grant. Congratulations! (Nov. 2014)
- Dr. Pan gave a keynote speech at the ACM CIKM’14 workshop on Interactive Mining of Big Data. (Nov. 2014)