1. K-State home
  2. »Division of Communications and Marketing
  3. »K-State Today
  4. »Statistics students and faculty present research at conferences

K-State Today

August 21, 2018

Statistics students and faculty present research at conferences

By Mike Higgins

Four graduate students and one faculty member from the College of Arts and Sciences' statistics department presented original research at recent conferences.

Graduate students Jianmei Luo, Behnaz Moradijamei and Haoyu Zhang and assistant professor Trevor Hefley presented at the Joint Statistical Meetings hosted in Vancouver, British Columbia, Canada, from July 28 to Aug. 2. Graduate student Kessinee Chitakasempornkul presented at the International Biometric Conference hosted in Barcelona, Spain, July 8-13.

Luo presented work on threshold clustering, a clustering method designed for big data applications in which clusters of a pre-specified minimum size are formed to make the within-cluster dissimilarity is small. Luo introduced iterative hybridized threshold clustering in which threshold clustering is used as a preprocessing step to aid in the scalability and robustness of other popular clustering algorithms. Simulation results show that iterative hybridized threshold clustering combined with popular clustering methods decreases the computer run time and memory usage of the popular clustering algorithms while still preserving their original performance. Luo is supervised by Michael Higgins, assistant professor of statistics.

Moradijamei presented work on renewal non-backtracking random walks, which are used to improve the performance of algorithms designed to detect communities such as those on social media platforms. Moradijamei presented results that suggest renewal non-backtracking random walks can improve existing community detection algorithms and the methodology can be applied efficiently to community detection problems with millions of individuals. Moradijamei also is supervised by Higgins.

Zhang presented work on novel Bayesian methodology to model the distribution of species across time and space. Mapping the spatial and temporal distribution of species is important for conserving the world's biodiversity. In collaboration with the Upper Midwest Environmental Sciences Center, Zhang develops a new approach that can be used to capture the interaction of multiple species which allows for multispecies data sources to be analyzed simultaneously. The end result is more accurate maps of where species occur. Zhang is supervised by Hefley.

Hefley presented work on using partial differential equations to develop a modeling framework that can be used to forecast the spread of white-nose syndrome for multiple species of bats across time and space. In addition, Hefley's framework can be used to estimate the location and time that white-nose syndrome was introduced into North America. Hefley's work will be used to inform disease surveillance efforts for white-nose syndrome.

Chitakasempornkul presented work using hierarchical Bayesian structural equation modeling to better assess animal production systems. In particular, Chitakasempornkul's methodology improves the ability to capture heterogeneous relationships between performance outcomes in animal production systems. Chitakasempornkul is supervised by Nora Bello, associate professor of statistics.

Luo and Chitakasempornkul received Graduate Student Council Travel awards to attend their meetings. Moradijamei, Zhang, and Chitakasempornkul received Arts and Sciences Research Travel Awards to attend their meetings. Additionally, Bello received a Faculty Development Award to attend the International Biometric Conference. 

Statisticians at K-State have been dedicated to the development, understanding, abstraction and communication of statistical principles for data analysis to Kansas State researchers for 55 years. The statistics department's long tradition of expertise in the design and analysis of experiments has been a critical component in many Kansas State research programs, helping them achieve national prominence.