Diverse Subset

Diverse Subset is used to rank entries in a database based on their distance from each other. The ranking order is suitable for extracting a subset of the database comprising entries which are farthest from each other by some diversity metric.

Options

Molecule Field
The database field containing the molecules
Subset Size
A value of zero means that no filtering is done. A positive value (smaller than the number of entries) means that a subset of the requested size will be produced.
Unit
Unit for training set size. Mols equals absolute number of molecules. % equals percentage of input data table.
Method
Method used to calculate the diverse Subset. The options for the selected method have to be set in the corresponding tab. Clustering uses k-means based on the fingerprint selected in that tab. The cluster centers are written out as the subset.
Descriptors
Select all numerical descriptor fields that should be included to calculate the diversity ranking.
Normalize Descriptors to unit variance.
Descriptor values will be normalized to unit variance. Strongly recommended!
Fingerprint
Fingerprint used to calculate the diversity.
Metric
Select Metric
Maximum Iterations
Maximum number of iterations for k-Means in Clustering mode.

Input Ports

Icon
Molecules

Output Ports

Icon
Diverse Subset

Popular Predecessors

Popular Successors

Views

Diverse Subset
Summary

Workflows

Links

Developers

You want to see the source code for this node? Click the following button and we’ll use our super-powers to find it for you.