BHI Colloquium: Tonglin Zhang, Ph.D. & Baijian "Justin" Yang, Ph.D.
A Few Statistical Considerations in Big Data Research
We are living in the Big Data era. It is estimated that in 2018, the data collected by just the IoT devices will reach 400 Zettabytes (10^21) in size. In this situation, traditional statistical approaches and algorithms are impossible to apply because of the memory and computational efficiency barriers. To overcome the difficulty, we have investigated properties of the well-known concept of sufficient statistics under a variety of statistical models. We have identified a number of cases where the size of sufficient statistics is not large even if the size of big data is massive. In the case when the size of sufficient statistics is also large, we modify the previous concept of sufficient statistics to a new concept called working sufficient statistics such that the size of working sufficient statistics is still small. Therefore, the implementation of sufficient statistics or working sufficient statistics can be used to efficiently overcome the memory or computational efficiency barriers in big data analysis. In this talk, we will present two examples based on the concept of sufficient statistics. We will also introduce possible extensions of our approaches based on the concept of working sufficient statistics.
About Tonglin Zhang & Justin Yang
Dr. Tonglin Zhang is currently an associate professor of statistics, Purdue University. He got his Ph.D. in statistics from the University of Michigan 2002. His research interests include mathematical statistics, spatial statistics for epidemiology, linear and generalized linear mixed effect models, and big data. He is currently an elected member of International Statistical Institute (ISI).
Dr. Baijian "Justin" Yang is currently an Associate Professor of Computer and Information Technology at Purdue University. His recent research interests are in cybersecurity, AI, and big data analysis. He was a member of IEEE Cybersecurity Initiative Steering Committee and the project lead of IEEE Try-CybSI from 2015 to 2017. He is currently serving as the Faculty Champion for the Polytechnic Research Impact Areas in Holistic Safety & Security
- Friday November 02, 2018 12:00 PM
- Friday November 02, 2018 01:00 PM
- IT 252
- Robyn Hart
- Contact Email