Randomized Parameter Search

Performs a randomized parameter search returning scores and best model found. Uses table as a “hypercube” of parameters to sample from. Ie. all combinations of parameters from all columns have equal chance of beeing sampled. Use masked values for parameter/columns with different lengths.

Definition

Input ports

in-model model

in-model

parameter space table

param-space

X table

X

Y table

Y

cross-validation 0 - 1, [(table,table)]

cross-validation

Output ports

results table

results

parameters table

parameters

out-model model

out-model

Configuration

Cross validation splits (cv)

Number of fold in the default K-Fold cross validation. Ignored when cross-validation port is given

iterations (n_iter)

Number of randomized searches done

Implementation

class node_paramsearch.ParameterSearch_Randomized