Two-Stage Sensitivity-Based Group Screening in Computer Experiments

2012/11/01 by Hyejung Moon, Angela M. Dean, Angela Dean +1 · 74 citations
Computer Science · Decision Sciences · Engineering · #Advanced Multi-Objective Optimization Algorithms #Algorithm #Artificial intelligence #Benchmark (surveying) #Code (set theory) #Computer science #Data mining #Electronic engineering #Engineering #Group (periodic table) #Machine learning #Noise (video) #Optimal Experimental Design Methods #Probabilistic and Robust Engineering Design #Programming language #Sensitivity (control systems) #Set (abstract data type) #Stage (stratigraphy)

paper · doi:10.1080/00401706.2012.725994

published in Technometrics 54(4), 376-387 (Taylor & Francis)

openalex publication_date 2012/11/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/29

Abstract

Sophisticated computer codes that implement mathematical models of physical processes can involve large numbers of inputs, and screening to determine the most active inputs is critical for understanding the input-output relationship. This article presents a new two-stage group screening methodology for identifying active inputs. In Stage 1, groups of inputs showing low activity are screened out; in Stage 2, individual inputs from the active groups are identified. Inputs are evaluated through their estimated total (effect) sensitivity indices (TSIs), which are compared with a benchmark null TSI distribution created from added low noise inputs. Examples show that, compared with other procedures, the proposed method provides more consistent and accurate results for high-dimensional screening. Additional details and computer code are provided in supplementary materials available online.

Citations

• The Detection of Defective Members of Large Populations
• Choosing the Sample Size of a Computer Experiment: A Practical Guide

Cited by

• Grouped Latin Hypercube Designs with Controlled Correlations
• Global Sensitivity Analysis with Dependence Measures
• Machine learning ranking of plausible (un)explored synergistic gene combinations using sensitivity indices of time series measurements of Wnt signaling pathway
• Efficient optimization of expensive black-box simulators via marginal means, with application to neutrino detector design
• Linear screening for high-dimensional computer experiments
• Unveil Sources of Uncertainty: Feature Contribution to Conformal Prediction Intervals
• Sparse polynomial chaos expansions via compressed sensing and D-optimal design
• Hilbert-Schmidt and Sobol sensitivity indices for static and time series Wnt signaling measurements in colorectal cancer - part A. [europepmc]

Related

• Design of Experiments for Screening
• Linear screening for high-dimensional computer experiments
• Choosing the Sample Size of a Computer Experiment: A Practical Guide
• Screening the Discrepancy Function of a Computer Model
• Factorial Sampling Plans for Preliminary Computational Experiments