DM2 Lab: Data Mining towards Decision Making

"Build trustworthy intelligent systems to support decision-making process with rich knowledge in massive behavior data." -- dm_square

Data-driven User Behavior Modeling: Understand behavior intention, model social and spatio-temporal context, represent unstructured behavior content, and consolidate into real use such as recommender systems, anti-fraud systems, and event monitoring systems.



Graduate Students

Qingkai Zeng: PhD student (2018)
Text mining: Data-driven syllabus design
Q: Given all data science papers, can machines automatically read the papers and design syllabus for the Data Science class? The methodology would be extended to any research field besides Data Science.
Haiqiao Zhang: PhD student (2016), co-advised with Dr. Yiyu Shi
Text mining: Concept hierarchy construction
Q: Given scientific papers (computer science, biological science, etc.), can we find concept terms and construct multifaceted concept hierarchy with little supervision?
Xueying Wang: PhD student (2016)
Text mining: Conditional/temporal fact extraction
Q: Given text documents (news, papers), can we extract (entity, attribute, value)-tuples, called EAV-tuples, and their conditions (e.g., from when to when, under what temperature)?
Daheng Wang: PhD student (2016), co-advised with Dr. Nitesh Chawla
Representation learning: Mining dynamic heterogeneous social media data
Q: Given heterogeneous signals/structures such as social networks, topics/events, spatio-temporal dimensions, and sentiments, can we learn low-dimensional representations for efficient social media mining?
Tong Zhao: Master student (2017)
Suspicious behavior detection: Fraud and bullying prevention
Q: Can we use data science to protect users on web platforms? We are delivering actionable solutions!


Zachary Eberhart
Graduate research: Sept.-Nov. 2017
Representation learning
Qi Li
Graduate research: Sept. 2017-Jan. 2018
Network science and text mining

Publications: Text Mining

Publications: Representation Learning

Publications: Suspicious Behavior Detection