Model formula.

Data frame for formula.

Matrix or dataframe containing the independent variables.

Vector containing the response variable, or, in the case of multiple responses, a matrix or dataframe whose columns are the values for each response.

Index vector specifying which cases to use, i.e., which rows in x to use. Default is NULL, meaning all.

Case weights. Default is NULL, meaning no case weights. If specified, weights must have length equal to nrow(x) before applying subset. Zero weights are converted to a very small nonzero value. In the current implementation, building models with weights can be slow.

Response weights. Default is NULL, meaning no response weights. If specified, wp must have an element for each column of y (after factors in y, if any, have been expanded). Zero weights are converted to a very small nonzero value.

NA action. Default is na.fail, and only na.fail is supported.

Offset term passed from the formula in earth.formula.

Default is FALSE. Set to TRUE to retain the following in the returned value: x and y (or data), subset, and weights. The function update.earth and friends will use these if present instead of searching for them in the environment at the time update.earth is invoked.
When the nfold argument is used with keepxy=TRUE, earth keeps more data and calls predict.earth multiple times to generate cv.oof.rsq.tab and cv.infold.rsq.tab (see the cv. arguments in earth.object). It therefore makes cross-validation significantly slower.

Trace earth's execution. Values:
0 (default) no tracing
.3 variance model (the varmod.method arg)
.5 cross validation (the nfold arg)
1 overview
2 forward pass
3 pruning
4 model mats summary, pruning details
5 full model mats, internal details of operation

NULL (default) or a list of arguments to pass on to glm. See the documentation of glm for a description of these arguments See “Generalized linear models” in the vignette. Example:
earth(survived~., data=etitanic, degree=2, glm=list(family=binomial))

The following arguments are for the forward pass.

Maximum degree of interaction (Friedman's mi). Default is 1, meaning build an additive model (i.e., no interaction terms).

Generalized Cross Validation (GCV) penalty per knot. Default is if(degree>1) 3 else 2. Simulation studies suggest values in the range of about 2 to 4. The FAQ section in the vignette has some information on GCVs.
Special values (for use by knowledgeable users): The value 0 penalizes only terms, not knots. The value -1 means no penalty, so GCV = RSS/n.

Maximum number of model terms before pruning, i.e., the maximum number of terms created by the forward pass. Includes the intercept.
The actual number of terms created by the forward pass will often be less than nk because of other stopping conditions. See “Termination conditions for the forward pass” in the vignette.
The default is semi-automatically calculated from the number of predictors but may need adjusting.

Forward stepping threshold. Default is 0.001. This is one of the arguments used to decide when forward stepping should terminate: the forward pass terminates if adding a term changes RSq by less than thresh. See “Termination conditions for the forward pass” in the vignette.

Minimum number of observations between knots. (This increases resistance to runs of correlated noise in the input data.)
The default minspan=0 is treated specially and means calculate the minspan internally, as per Friedman's MARS paper section 3.8 with alpha = 0.05. Set trace>=2 to see the calculated value.
Use minspan=1 and endspan=1 to consider all x values.
Negative values of minspan specify the maximum number of knots per predictor. These will be equally spaced. For example, minspan=-3 allows three evenly spaced knots for each predictor. As always, knots that fall in the end zones specified by endspan will be ignored.

Minimum number of observations before the first and after the final knot.
The default endspan=0 is treated specially and means calculate the endspan internally, as per the MARS paper equation 45 with alpha = 0.05. Set trace>=2 to see the calculated value.
Be wary of reducing endspan, especially if you plan to make predictions beyond or near the limits of the training data. Overfitting near the edges of training data is much more likely with a small endspan. The model's RSq and GRSq won't indicate when this overfitting is occurring. (A plotmo plot can help: look for sharp hinges at the edges of the data). See also the Adjust.endspan argument.

Penalty for adding a new variable in the forward pass (Friedman's gamma, equation 74 in the MARS paper). Default is 0, meaning no penalty for adding a new variable. Useful non-zero values typically range from about 0.01 to 0.2 and sometimes higher — you will need to experiment.
A word of explanation. With the default newvar.penalty=0, if two variables have nearly the same effect (e.g. they are collinear), at any step in the forward pass earth will arbitrarily select one or the other (depending on noise in the sample). Both variables can appear in the final model, complicating model interpretation. On the other hand with a non-zero newvar.penalty, the forward pass will be reluctant to add a new variable — it will rather try to use a variable already in the model, if that does not affect RSq too much. The resulting final model may be easier to interpret, if you are lucky. There will often be a small performance hit (a worse GCV).

Maximum number of parent terms considered at each step of the forward pass. (This speeds up the forward pass. See the Fast MARS paper section 3.0.)
Default is 20. A value of 0 is treated specially (as being equivalent to infinity), meaning no Fast MARS. Typical values, apart from 0, are 20, 10, or 5.
In general, with a lower fast.k (say 5), earth is faster; with a higher fast.k, or with fast.k disabled (set to 0), earth builds a better model. However, because of random variation this general rule often doesn't apply.

Fast MARS ageing coefficient, as described in the Fast MARS paper section 3.1. Default is 1. A value of 0 sometimes gives better results.

Index vector specifying which predictors should enter linearly, as in lm. The default is FALSE, meaning predictors enter in the standard MARS fashion, i.e., in hinge functions.

The linpreds argument does not specify that a predictor must enter the model; only that if it enters, it enters linearly. See “The linpreds argument” in the vignette.
See also the Auto.linpreds argument below (which describes how earth will automatically treat a predictor as linear under certain conditions).

Details: A predictor's index in linpreds is the column number in the input matrix x (after factors have been expanded).
linpreds=TRUE makes all predictors enter linearly (the TRUE gets recycled).
linpreds may be a character vector e.g. linpreds=c("wind", "vis"). Note: grep is used for matching. Thus "wind" will match all variables that have "wind" in their names. Use "^wind$" to match only the variable named "wind".

Function specifying which predictors can interact and how. Default is NULL, meaning all standard MARS terms are allowed.
During the forward pass, earth calls the allowed function before considering a term for inclusion; the term can go into the model only if the allowed function returns TRUE. See “The allowed argument” in the vignette.

The following arguments are for the pruning pass.

Pruning method. One of: backward none exhaustive forward seqrep cv.
Default is "backward".
Specify pmethod="cv" to use cross-validation to select the number of terms. This selects the number of terms that gives the maximum mean out-of-fold RSq on the fold models. Requires the nfold argument.
Use "none" to retain all the terms created by the forward pass.
If y has multiple columns, then only "backward" or "none" is allowed.
Pruning can take a while if "exhaustive" is chosen and the model is big (more than about 30 terms). The current version of the leaps package used during pruning does not allow user interrupts (i.e., you have to kill your R session to interrupt; in Windows use the Task Manager or from the command line use taskkill).

Maximum number of terms (including intercept) in the pruned model. Default is NULL, meaning all terms created by the forward pass (but typically not all terms will remain after pruning). Use this to enforce an upper bound on the model size (that is less than nk), or to reduce exhaustive search time with pmethod="exhaustive".

The following arguments are for cross validation.

Number of cross-validation folds. Default is 0, no cross validation. If greater than 1, earth first builds a standard model as usual with all the data. It then builds nfold cross-validated models, measuring R-Squared on the out-of-fold (left out) data each time. The final cross validation R-Squared (CVRSq) is the mean of these out-of-fold R-Squareds.
The above process of building nfold models is repeated ncross times (by default, once). Use trace=.5 to trace cross-validation.
Further statistics are calculated if keepxy=TRUE or if a binomial or poisson model (specified with the glm argument). See “Cross validation” in the vignette.

Only applies if nfold>1. Number of cross-validations. Each cross-validation has nfold folds. Default 1.

Only applies if nfold>1. Default is TRUE. Stratify the cross-validation samples so that an approximately equal number of cases with a non-zero response occur in each cross validation subset. So if the response y is logical, the TRUEs will be spread evenly across folds. And if the response is a multilevel factor, there will be an approximately equal number of each factor level in each fold (because a multilevel factor response gets expanded to columns of zeros and ones, see “Factors” in the vignette). We say “approximately equal” because the number of occurrences of a factor level may not be exactly divisible by the number of folds.

The following arguments are for variance models.

Construct a variance model. For details, see varmod and the vignette “Variance models in earth”. Use trace=.3 to trace construction of the variance model.
This argument requires nfold and ncross. (We suggest at least ncross=30 here to properly calculate the variance of the errors — although you can use a smaller value, say 3, for debugging.)
The varmod.method argument should be one of
"none" Default. Don't build a variance model.
"const" Assume homoscedastic errors.
"lm" Use lm to estimate standard deviation as a function of the predicted response.
"rlm" Use rlm.
"earth" Use earth.
"gam" Use gam. This will use either gam or the mgcv package, whichever is loaded.
"power" Estimate standard deviation as intercept + coef * predicted.response^exponent, where intercept, coef, and exponent will be estimated by nls. This is equivalent to varmod.method="lm" except that exponent is automatically estimated instead of being held at the value set by the varmod.exponent argument.
"power0" Same as "power" but no intercept (offset) term.
"x.lm", "x.rlm", "x.earth", "x.gam" Like the similarly named options above, but estimate standard deviation by regressing on the predictors x (instead of the predicted response). A current implementation restriction is that "x.gam" allows only models with one predictor (x must have only one column).

Power transform applied to the rhs before regressing the absolute residuals with the specified varmod.method. Default is 1.
For example, with varmod.method="lm", if you expect the standard deviance to increase linearly with the mean response, use varmod.exponent=1. If you expect the standard deviance to increase with the square root of the mean response, use varmod.exponent=.5 (where negative response values will be treated as 0, and you will get an error message if more than 20% of them are negative).

Convergence criterion for the Iteratively Reweighted Least Squares used when creating the variance model.
Iterations stop when the mean value of the coefficients of the residual model change by less than varmod.conv percent. Default is 1 percent.
Negative values force the specified number of iterations, e.g. varmod.conv=-2 means iterate twice.
Positive values are ignored for varmod="const" and also currently ignored for varmod="earth" (these are iterated just once, the same as using varmod.conv=-1).

The estimated standard deviation of the main model errors is forced to be at least a small positive value, which we call min.sd. This prevents negative or absurdly small estimated standard deviations. Clamping takes place in predict.varmod, which is called by predict.earth when estimating prediction intervals. The value of min.sd is determined when building the variance model as min.sd = varmod.clamp * mean(sd(training.residuals)). The default varmod.clamp is 0.1.

Only applies when varmod.method="earth" or "x.earth". This is the minspan used in the internal call to earth when creating the variance model (not the main earth model).
Default is -3, i.e., three evenly spaced knots per predictor. Residuals tend to be very noisy, and allowing only this small number of knots helps prevent overfitting.

The following arguments are for internal or advanced use.

Earth object to be updated, for use by update.earth.

Scale the response internally in the forward pass. Scaling here means subtract the mean and divide by the standard deviation.
For single-response models, the default is Scale.y = TRUE. Scaling is invisible to the user, up to numerical differences, but does provide better numeric stability.
For multiple-response models, the default is FALSE. If Scale.y is set TRUE, each column of the response is independently scaled. This can prevent one response from “overwhelming” the others, and earth typically generates a different set of hinge functions.

In interaction terms, endspan gets multiplied by this value. This reduces the possibility of an overfitted interaction term supported by just a few cases on the boundary of the predictor space (as sometimes seen in our simulation studies).
The default is 2. Use Adjust.endspan=1 for compatibility with old versions of earth.

Default is TRUE, which works as follows (see example):
At any step in the forward pass, if earth discovers that the best knot for the best predictor is at the predictor minimum (in the training data), then earth adds the predictor to the model as a linear “basis function” (with no hinge). Compare the following basis functions (printed in bold) for an example such predictor x:
Auto.linpreds=TRUE (default): x
Auto.linpreds=FALSE: max(x-99, 0) where 99 is the minimum x in the training data.
Using Auto.linpreds=FALSE always forces a knot, even when the knot is at the minimum value of the variable. This ensures that the basis functions are always expressed as hinge functions (and will always be non-negative).
Note that Auto.linpreds affects only how the model behaves outside the training data. Thus predict.earth will make the same predictions from the training data, regardless of whether the earth model was built with Auto.linpreds set TRUE or FALSE (up to possible differences in the size of the model caused by different GCVs because of the different forms of the terms).

Default is FALSE. For testing the weights argument. Force use of the code for handling weights in the earth code, even if weights=NULL or all the weights are the same. This will not necessarily generate an identical model, primarily because the non-weighted code requires some tests for numerical stability that can sometimes affect knot selection.

Default is TRUE. Using the “beta cache” takes a little more memory but is faster (by 20% and often much more for large models). The beta cache uses nk * nk * ncol(x) * sizeof(double) bytes. (The beta cache is an innovation in this implementation of MARS and does not appear in Friedman's papers. It is not related to the fast.beta argument. Certain regression coefficients in the forward pass can be saved and re-used, thus saving recalculation time.)

Default is FALSE. This argument pertains to subset evaluation in the pruning pass. By default, if y has a single column then earth calls the leaps routines; if y has multiple columns then earth calls EvalSubsetsUsingXtx. The leaps routines are numerically more stable but do not support multiple responses (leaps is based on the QR decomposition and EvalSubsetsUsingXtx is based on the inverse of X'X). Setting Force.xtx.prune=TRUE forces use of EvalSubsetsUsingXtx, even if y has a single column.

Default is TRUE unless the model has more than 100 thousand cases. The leverages are the diagonal hat values for the linear regression of y on bx. (The leverages are needed only for certain model checks, for example when plotres is called with versus=4).
Details: This argument was introduced to reduce peak memory usage. When n >> p, memory use peaks when earth is calculating the leverages.

Default 1e-10. Applies only when pmethod="exhaustive". If the reciprocal of the condition number of bx is less than Exhaustive.tol, earth forces pmethod="backward". See “XHAUST returned error code -999” in the vignette.

Dots are passed on to earth.fit.