Joining Forces from Data Mining and Visual Analytics for Big Data
Two different approaches exist for analyzing big data: data mining/machine learning methods and visual analytics approaches. Each of them has its own pros and cons, and by utilizing both, we can better solve problems. As ways to achieve this claim, this talk will cover the following.
(1) I will present a system called FODAVA Testbed, which integrates various dimension reduction and clustering methods with visual analytics for general high-dimensional data including text, image, bioinformatics data, etc.
(2) In the context of document topic modeling and clustering, I will present another system called UTOPIAN, a user-driven topic modeling based on interactive nonnegative matrix factorization. (http://www.youtube.com/watch?v=du6_s6hcaRA)
(3) I~@~Yll talk about several research ideas and open questions for scaling up the computational methods in big data visual analytics.
Jaegul Choo is a Research Scientist in CSE, working with Dr. Haesun Park. Under her supervision, Jaegul received a Ph.D in CSE in 2013. He~@~Ys broadlyy interested in combining data mining techniques with visual analytics in order to leverage both computational and human capabilities for real-world data analysis problems.