Congratulations to Ph.D. Student Ji Hwan Oh



Ph.D. Student Ji Hwan Oh wins first place award at the ICSA Midwest Chapter meeting’s student poster competition on October 26, 2015.

Congratulations to Ph.D. student Ji Hwan Oh (advisor Dr. Hyonho Chun), who was recently awarded the first place prize for the student poster competition at the International Chinese Statistical Association (ICSA) Midwest Chapter meeting for his poster titled Kernal Partial Correlation: A Novel Approach to Capturing Conditional Independence in Graphical Models for Noisy Data co-authored with fellow Ph.D. candidate Faye Zheng and Professor R.W. Doerge.

This award is bestowed and evaluated to the students with the best presence and quality of statistical significance, novelty of the technical content, accessibility to a broad audience, accuracy of language, layout of content, readability of the presentation, and graphics. The top three posters were awarded.

Ji Hwan’s poster focused on graphical models. Graphical models capture the conditional independence structure among random variables via the existence of edges among vertices. Under the assumption of multivariate normality, the detection of conditional independence between two variables given other variables is equivalent to the identification of a zero partial correlation coefficient. They propose a new measure, kernel partial correlation, which is estimated using a combination of two statistical methods. First, a nonparametric regression is employed for conditioning, and then the nonparametric association is assessed to detect independence. Since their approach does not rely on heavy distributional assumptions, it is appropriate for situations where the assumption of a multivariate Gaussian distribution fails to hold, especially when the noise is high. Upon comparisons to existing approaches, their method outperforms when it is applied to simulated data as well as real data from single-cell RNA-sequencing experiments.

Join us in congratulating Ji Hwan on his award.