21,000 datasets and 100,000 machine learning tasks are accessible through this R interface to the OpenML platform. Users can programmatically search metadata, download data objects, and upload experimental results using standardized API functions.
Use Cases
- Filter for specific dataset characteristics using the number.of.instances and number.of.missing.values columns in the metadata table
- Benchmark R-based models against existing community results by comparing predictive.accuracy metrics for a given task.id
- Programmatically retrieve data for specific domains by searching the tag column in the listOMLDataSets output
Strengths
- listOMLDataSets returns a table of 21,000+ entries with columns for number.of.features and number.of.classes
- getOMLTask provides access to 100,000+ tasks containing target.feature labels and estimation.procedure details
- listOMLFlows provides a registry of machine learning algorithms and their associated external.version strings