Constructing Efficient Experimental Designs for Generalized Linear Models

Saleh, Moein; Pan, Rong

doi:10.1080/03610918.2014.927486

Cited by 3 publications

(1 citation statement)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are many approaches that address individually the optimal DoE problem and missing data imputations but there are none to our knowledge that solve them simultaneously. Methods specifically developed for DOE that apply to various optimality criteria include heuristic algorithms such as genetic algorithm [21], [22] and Fedorov's exchange [23] or convex relaxations [24]; some approaches provide better analytical guarantees for specific criteria such as T − and Doptimality [25]- [27], or A-optimality [28]- [31]. On the other hand imputing values to missing-data problems are generally addressed by using surrogates such as mean or mode of the data [32]- [34]; or estimating missing values using maximum likelihood methods [35].…”

Section: Introductionmentioning

confidence: 99%

Design of Experiments with Imputable Feature Data: An Entropy-Based Approach

Velicheti¹,

Srivastava²,

Salapaka³

2021

Preprint

View full text Add to dashboard Cite

Tactical selection of experiments to estimate an underlying model is an innate task across various fields. Since each experiment has costs associated with it, selecting statistically significant experiments becomes necessary. Classic linear experimental design deals with experiment selection so as to minimize (functions of) variance in estimation of regression parameter. Typically, standard algorithms for solving this problem assume that data associated with each experiment is fully known. This isn't often true since missing data is a common problem. For instance, remote sensors often miss data due to poor connection. Hence experiment selection under such scenarios is a widespread but challenging task. Though decoupling the tasks and using standard data imputation methods like matrix completion followed by experiment selection might seem a way forward, they perform sub-optimally since the tasks are naturally interdependent. Standard design of experiments is an NP hard problem, and the additional objective of imputing for missing data amplifies the computational complexity. In this paper, we propose a maximum-entropy-principle based framework that simultaneously addresses the problem of design of experiments as well as the imputation of missing data. Our algorithm exploits homotopy from a suitably chosen convex function to the non-convex cost function; hence avoiding poor local minima. Further, our proposed framework is flexible to incorporate additional application specific constraints. Simulations on various datasets show improvement in the cost value by over 60% in comparison to benchmark algorithms applied sequentially to the imputation and experiment selection problems.

show abstract