Abstract
Searching active factors in supersaturated screening experiments is a difficult task. A Stepwise selection procedure is known to be ineffective because it tends to select too many inactive factors and miss some active factors. Several methods have been proposed in the literature in recent years, which are effective in some aspects. In this article, we introduce a new strategy of searching active factors in supersaturated screening experiments based on the idea of staged dimensionality reduction. Simulation and real data studies demonstrate that this strategy is quite effective. It is also easy to understand and implement.
Additional information
Notes on contributors
Xuan Lu
Dr. Lu is a professor in the Department of Mathematical Science. His email address is [email protected].
Xi Wu
Ms. Wu is a graduate student in the Department of Mathematical Science. Her email address is [email protected].