Schemashyperparameters/v1

Hyperparameters

Component

Language-independent algorithm hyperparameters: an algorithm name and an algorithm-specific `hyperparameters` object, validated per-algorithm against schema.rtemis.org/hyperparameters/<algorithm>/v1. Mirrors the `{algorithm, hyperparameters}` wire format consumed by `.list_to_Hyperparameters`.

Raw JSONv1Unknown properties rejected

Variants

The value of algorithm selects which schema applies to hyperparameters.

GLMGLMHyperparameters
GAMGAMHyperparameters
GLMNETGLMNETHyperparameters
GLMTreeGLMTreeHyperparameters
SPLSSPLSHyperparameters
MARSMARSHyperparameters
LinearSVMLinearSVMHyperparameters
RadialSVMRadialSVMHyperparameters
CARTCARTHyperparameters
RangerRangerHyperparameters
LightCARTLightCARTHyperparameters
LightRFLightRFHyperparameters
LightGBMLightGBMHyperparameters
LightRuleFitLightRuleFitHyperparameters
BARTBARTHyperparameters
IsotonicIsotonicHyperparameters
MLPMLPHyperparameters
TabNetTabNetHyperparameters
KNNKNNHyperparameters
HALHALHyperparameters
MonotonicHALMonotonicHALHyperparameters
LINADLINADHyperparameters
LINADForestLINADForestHyperparameters
NNLSNNLSHyperparameters
SuperLearnerSuperLearnerHyperparameters
ModalityStackingModalityStackingHyperparameters
ConditionalSuperLearnerConditionalSuperLearnerHyperparameters

Properties

algorithmrequired

string

Supervised-learning algorithm name (matches `setup_<algorithm>`).

one of"GLM""GAM""GLMNET""GLMTree""SPLS""MARS""LinearSVM""RadialSVM""CART""Ranger""LightCART""LightRF""LightGBM""LightRuleFit""BART""Isotonic""MLP""TabNet""KNN""HAL""MonotonicHAL""LINAD""LINADForest""NNLS""SuperLearner""ModalityStacking""ConditionalSuperLearner"

hyperparametersrequired

object

Variant-specific parameters. Validated per `algorithm` below.

Relationships

Used by

supervisedconditionalsuperlearnermodalitystackingsuperlearner