MARSHyperparameters

Properties

degree

integer | object≥ 1

Maximum degree of interaction. 1 builds an additive model with no interaction terms.

penalty

number | object | null≥ -1

Generalized Cross Validation penalty per knot. NULL uses 3 when @degree is greater than 1 and 2 otherwise. 0 penalizes terms but not knots, and -1 removes the penalty.

nk

integer | object | null≥ 1

Maximum number of terms, including the intercept, created by the forward pass. NULL lets earth derive it from the number of features.

nprune

integer | object | null≥ 1

Maximum number of terms, including the intercept, retained after pruning. NULL keeps every term the forward pass created.

thresh

number | object≥ 0< 1

Forward pass stopping threshold: stop once adding a term changes R-squared by less than this.

minspan

integer | object

Minimum number of observations between knots. 0 derives the value internally, and a negative value instead sets the maximum number of equally spaced knots per feature.

endspan

integer | object≥ 0

Minimum number of observations before the first and after the final knot. 0 derives the value internally.

newvar_penalty

number | object≥ 0

Penalty for adding a feature not already in the model during the forward pass. 0 applies no penalty; useful values typically range from 0.01 to 0.2.

fast_k

integer | object≥ 0

Maximum number of parent terms considered at each step of the forward pass. 0 disables Fast MARS, which is slower but builds a better model.

pmethod

string

Pruning method. "cv" selects the number of terms by cross-validation and requires @nfold. Multiclass classification allows only "backward" and "none".

one of"backward""none""exhaustive""forward""seqrep""cv"

nfold

integer≥ 0

Number of cross-validation folds used to estimate out-of-fold R-squared. 0 disables cross-validation. Cannot exceed the number of cases in the training data.

ncross

integer≥ 1

Number of times the @nfold cross-validation is repeated. Applies only when @nfold is greater than 1.

stratify

boolean

Stratify the cross-validation folds on the outcome. Applies only when @nfold is greater than 1.

fast_beta

number≥ 0≤ 1

Fast MARS aging coefficient. 0 sometimes gives better results.

ifw

boolean | object

Inverse Frequency Weighting in classification.

Relationships