Congratulations to biostatistics alumnus Ryan Peterson (19PhD, 16MS), whose first-authored paper “Ordered quantile normalization: A semiparametric transformation built for the cross-validation era” has been selected to receive the 2020 Journal of Applied Statistics Best Paper Prize. The winning article receives a £500 prize (or USD equivalent) and the paper will be made free to view for the following year.
“Ryan developed this methodology, as well as the R package for its implementation, while he was a doctoral student in our program,” says Joe Cavanaugh, professor and head of biostatistics and co-author of the paper. “Although the paper was only published a year ago, according to Google Scholar, it has already been cited over 150 times.”
Peterson is currently an assistant professor of biostatistics and informatics at the Colorado School of Public Health.
The statement from the JAS Award Selection Committee, which summarizes the contribution, appears below.
The authors of this article proposed a new transformation, called Ordered Quantile (ORQ) normalization, to produce normally distributed data from data that follow any arbitrary distribution. Extensive simulation studies were conducted for the cases where the data were generated from known distributions of the asymmetry, bimodal and heavy-tailed types. The effectiveness of the ORQ technique was compared with other popular normalization methods. The proposed normalization transformation guarantees, in the absence of ties, to produce normally distributed transformed data that is related one-to-one with the original data. The proposed method is also incorporated in an R package (bestNormalize) to facilitate its use. Such a technique is very useful, especially in an unsupervised machine learning framework, which often requires normally distributed data.