AI Paper Atlas

Research map / Probabilistic and scientific modeling / Uncertainty and Gaussian processes

Bayesian subset selection and variable importance for interpretable prediction and classification

Daniel R. Kowal

NeurIPS 2022 · Probabilistic and scientific modeling › Uncertainty and Gaussian processes · approach: Nonparametric and PAC-Bayes Bayesian learning

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use smallest acceptable subset S_small instead of S_min for selectionbbound(bayes) · simulation n=50, p=50, SNR=0.25TNR: 0.81 -> 0.98Improved
use S_small instead of S_minbbound(bayes) · simulation n=500, p=50, SNR=1TNR: 0.67 -> 0.99Improved
use S_small instead of S_minbbound(bayes) · simulation n=500, p=50, SNR=0.25TPR: 0.99 -> 0.86Hurt
S_small versus adaptive lasso for variable selectionadaptive lasso · simulation n=200, p=400, SNR=1TPR: 0.97 -> 0.93No clear effect
S_small versus adaptive lasso at small n, high SNRadaptive lasso · simulation n=50, p=50, SNR=1TPR: 0.54 -> 0.34Hurt
posterior HPD selection instead of S_smallS_small · simulation n=200, p=400, SNR=1TPR: 0.93 -> 0.75Hurt
forward selection instead of S_smallS_small · simulation n=50, p=50, SNR=0.25TNR: 0.98 -> 0.06Hurt
classical subset selection instead of S_smallS_small · simulation n=500, p=50, SNR=1TNR: 0.99 -> 0.83Hurt
use branch-and-bound search instead of lasso/forward/backward search pathsbbound(bayes) · simulation, acceptable family sizenumber of acceptable subsets: up to 10 times the number of acceptable subsetsImproved
branch-and-bound search on NC education out-of-sample splitsbbound(bayes) · NC education 20 splitsnumber of subsets: 10-25 times the number of subsetsImproved
reduce m_k from 100 to 15 (ablation)bbound(bayes) · NC education reading scoresnumber of acceptable subsets: 1183 -> 197No clear effect
increase epsilon from 0.1 to 0.2 (ablation)bbound(bayes) · NC education reading scoresnumber of acceptable subsets: 1183 -> 977No clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy