A Review of Global Sensitivity Analysis Methods and a comparative case study on Digit Classification

TL;DR

This paper reviews various global sensitivity analysis methods, compares them through a case study on MNIST digit classification, and proposes a methodology to evaluate their effectiveness in high-dimensional data contexts.

Contribution

It provides a comprehensive review of GSA methods, introduces a new evaluation methodology, and compares their performance on a real-world digit classification task.

Findings

Certain GSA methods effectively identify influential features in digit classification.

The proposed evaluation methodology offers a systematic way to compare GSA techniques.

Some methods outperform others in computational efficiency and accuracy.

Abstract

Global sensitivity analysis (GSA) aims to detect influential input factors that lead a model to arrive at a certain decision and is a significant approach for mitigating the computational burden of processing high dimensional data. In this paper, we provide a comprehensive review and a comparison on global sensitivity analysis methods. Additionally, we propose a methodology for evaluating the efficacy of these methods by conducting a case study on MNIST digit dataset. Our study goes through the underlying mechanism of widely used GSA methods and highlights their efficacy through a comprehensive methodology.