bestglm is a dataset or software package for best subset selection in generalized linear models and regression. It implements methods using information criteria or cross-validation, including the 'leaps' algorithm and complete enumeration, and also supports PCR and PLS. The work is authored by A.I. McLeod, Changjiang Xu, and Yuanhao Lai.
Use Cases
- Selecting optimal predictor subsets for regression models based on information criteria like AIC/BIC.
- Implementing principal component regression (PCR) or partial least squares (PLS) for dimensionality reduction.
- Applying the one-standard deviation rule for model tuning in conjunction with the 'caret' package.
- Comparing model selection algorithms like 'leaps' and complete enumeration for statistical analysis.
Strengths
- Implements established statistical algorithms from peer-reviewed literature, including Furnival and Wilson (1974) and Morgan and Tatar (1972).
- Supports multiple model selection and dimensionality reduction techniques (best subset GLM, PCR, PLS).
- Designed for integration with the widely-used 'caret' R package for machine learning.
Limitations
- Description metadata is limited; actual data quality requires manual inspection after download.
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count, file formats, and license information are unknown.
Provenance
- Source
- paperswithcode
- Collection Method
- Likely a software package or methodological dataset for statistical modeling.
- Time Range
- null
- Freshness
- Last update date is unknown; freshness unverified.
- Geography
- null