GEMSS: Subsampling Data for Gaussian Process Model Training

GEMSS-Driven Subsampling for Information Extraction and Redundancy Elimination is a methodological dataset and R package by Ming-Chung Chang, last updated in May 2026. The 6.6 MB resource includes PDF and ZIP files describing a subsampling approach to improve Gaussian process model accuracy in unexplored input regions. The method, Generalization Error Minimization in SubSampling (GEMSS), aims to identify informative data subsets while discarding redundant points.

Use Cases

• Training Gaussian process models on large datasets based on the described subsampling methodology.
• Improving predictive accuracy in unexplored regions of an input space based on the GEMSS approach.
• Identifying and eliminating redundant data points to address numerical instability in statistical models.
• Implementing the GEMSS subsampling technique using the provided R package.

Strengths

• Includes a 6.6 MB package with an accompanying R package for implementation.
• Methodology is supported by theoretical results and numerical studies across diverse scenarios, as stated in the description.
• Released under a permissive CC-BY-4.0 license.

Limitations

• Row count and column-level documentation are unknown, which limits suitability assessment.
• The dataset appears to be primarily methodological documentation and software; the nature and volume of any example data are unclear.
• Freshness should be verified as the last update timestamp is in the future (2026-05-07).

Provenance

Source

Ming-Chung Chang via figshare

Collection Method

Proposed subsampling methodology (GEMSS) for statistical modeling.

Freshness

Last updated 2026-05-07 17:59:44; freshness should be verified.