SELECT REPRESENTATIVE SUBSET

Selects a feature-space-diverse subset of the rows of a dataset (maximin / farthest-point sampling), for ML-friendly sampling of existing points.

When to use

Tagged: doe, sampling, subset, space_filling, maximin, ml, data_io.

Inputs

Label ID Type Default Required Description
Dataset dataset dataset Rows to select from.
Feature columns independents text Comma-separated feature column names defining the space.
Number of points (K or fraction) num_points text 0.2   Integer K, or a fraction 0<f<1 of the row count.

Outputs

Label ID Type Description
Selected selected dataset The chosen subset rows (tagged selected=yes).
Remaining remaining dataset The rows not chosen.
Selected Ids selected_ids dataset Ids of the chosen rows.

Disciplines

  • data.io
  • platform.data_management

Auto-generated from platform schema. Worker id: doe_subset_select. Schema hash: 076b2fb98189. Hand-curated docs in workerexamples/ override this page when present.