Academy of Mathematics and Systems Science, CAS Colloquia & Seminars
Speaker:
Frederick Kin Hing Phoa, Institute of Statistical Science, Academia Sinica
Inviter:
Title:
Designing Experiments with Unstructured Treatments for General Network Structures
Time & Venue:
2019.7.1 14:40-15:20 N205
Abstract:
Experiments on connected units are commonly conducted in many scientific fields. An experimental unit in these applications may connect with some others, and the treatment applied to a unit has an effect, called a network effect, towards the responses resulted in the neighboring units. Designing such experiments was rarely discussed in the literature. Parker, Gilmour, and Schormans (2017) initiated a study of A_s-optimal designs on connected experimental units with unstructured treatments, assuming that the network effects are unknown constants. This work investigates in a similar design problem but the network effects are assumed to be random effects, which lead to a property that the responses of two units are correlated if some neighbors of one unit and those of the other receive the same treatment. Alphabetical optimality criteria are considered for selecting good designs with high efficiency of estimating the treatment effects and/or high accuracy of predicting the network effects. We provide theoretical conditions for designs to be optimal and illustrate our theory with some numerical examples. This is a joint work with Dr. Ming-Chung Chang of National Central University and Ms. Jin-Wen Huang, National Tsing Hua University.