Inferring feature importance with uncertainties with application to large genotype data

Pål Vegard Johnsen, Inga Strümke, Mette Langaas, Andrew Thomas DeWan and Signe Riemer-Sørensen

PLOS Computational Biology, 2023, vol. 19, issue 3, 1-22

Abstract: Estimating feature importance, which is the contribution of a prediction or several predictions due to a feature, is an essential aspect of explaining data-based models. Besides explaining the model itself, an equally relevant question is which features are important in the underlying data generating process. We present a Shapley-value-based framework for inferring the importance of individual features, including uncertainty in the estimator. We build upon the recently published model-agnostic feature importance score of SAGE (Shapley additive global importance) and introduce Sub-SAGE. For tree-based models, it has the advantage that it can be estimated without computationally expensive resampling. We argue that for all model types the uncertainties in our Sub-SAGE estimator can be estimated using bootstrapping and demonstrate the approach for tree ensemble methods. The framework is exemplified on synthetic data as well as large genotype data for predicting feature importance with respect to obesity.Author summary: Artificial intelligence and machine learning have been increasingly popular tools for modelling complex relationships in medicine and genomics. For example a machine learning model for predicting the likelihood of a particular person developing some disease. The prediction model can for instance be based on genomics data, which consists of a large number of features for each single person. Such prediction models can be very complex and difficult to interpret, hence they are often denoted black-box models. However, to exploit the knowledge the prediction model has gained, we must be able to interpret it, and explain which features are important for the model, but also for the underlying data. We investigate a theoretical approach for extracting feature importance, even when the model input consists of many features. Lastly, we emphasize the need for estimating the uncertainty of the individual feature importance, and provide a bootstrap procedure for doing so.

Date: 2023
References: View complete reference list from CitEc
Citations:

Downloads: (external link)
https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1010963 (text/html)
https://journals.plos.org/ploscompbiol/article/fil ... 10963&type=printable (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:plo:pcbi00:1010963

DOI: 10.1371/journal.pcbi.1010963

Access Statistics for this article

More articles in PLOS Computational Biology from Public Library of Science
Bibliographic data for series maintained by ploscompbiol ().