VSURF: Variable Selection Procedure for High-Dimensional Data

A three-step variable selection procedure based on random forests, initially developed for high-dimensional data where variables exceed observations. The method is versatile for regression and supervised classification problems, as described by Genuer, Poggi, and Tuleau-Malot in a 2015 R Journal article. The package aims to eliminate irrelevant variables, select all response-related variables for interpretation, and refine the selection by removing redundancy for prediction.

Use Cases

• Perform initial variable elimination for high-dimensional datasets based on the described three-step procedure.
• Select variables related to a response for model interpretation purposes using the second-step method.
• Refine a variable set by eliminating redundancy for improved prediction performance as per the third step.
• Apply a versatile selection tool to regression problems as mentioned in the description.
• Apply a versatile selection tool to supervised classification problems as mentioned in the description.

Strengths

• Procedure is specifically designed for high-dimensional data where variables exceed observations.
• Method is described as versatile for both regression and supervised classification problems.
• Selection process is structured into three distinct steps for elimination, interpretation, and refinement.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

R Journal article by Genuer, R., Poggi, J.-M., and Tuleau-Malot, C. (2015).

Collection Method

Algorithmic procedure for variable selection.

Time Range

Publication date is 2015.

Freshness

Last updated date is unknown.

Geography

Spatial coverage is not specified.