A growing number of scholarly and professional fields utilize graphical modeling techniques to clearly visualize relationships and patterns among multiple, large sets of data. Yet, most of these modeling techniques are only accurate when working with data sets that are normally distributed—which is never a guarantee in practice.

Kuang-Yao Lee and his co-authors have developed a nonparametric graphical model for multivariate random functions that is flexible, accurate and computationally feasible when visualizing relationships and patterns among large sets of data that are not normally distributed.

Lee’s model can be integrated with existing modeling techniques, and researchers who frequently work with non-normally distributed data are encouraged to begin incorporating it to improve the accuracy of visualizations.