ml.svm {RGeostats}

R Documentation

Regression using SVM methodology

Description

Regression using SVM methodology

Usage

ml.svm(dbin, dbout, classification = FALSE, auxnames = NA, catnames = NA, 
tune = TRUE, kernel = c("radial"), cost = c(1), probability = FALSE, ...,
verbose = FALSE, radix = NA, modify.target = db.locmod())

Arguments

dbin	The db-class database containing the data
dbout	The db-class database containing the targets
classification	If TRUE, perform a classification (Interest variable should be categorical), otherwise perform a regression.
auxnames	List of optional auxiliary continuous covariables. See details for more information. If empty, only coordinates are used.
catnames	List of optional auxiliary categorical covariables. See details for more information. If empty, only coordinates are used.
tune	Set TRUE for tuning the model (find best hyper parameters).
kernel,cost,...	Arguments specific to SVM methodology. See svm for details.
probability	To be defined
verbose	Verbose flag.
radix	Radix of the name given to the variables storing the results in the target Db.
modify.target	Decides whether or not the newly created variables will have their locator defined or not. For more information, see db.locmod.

Details

The training is based on the following information from 'dbin':

• The coordinates

• The auxiliary continuous variable(s) (defined in 'auxnames')

• The auxiliary categorical variable name(s) (defined in 'catnames')

• The target variable (first variable with locator 'z')

The prediction is performed in 'dbout'. It requires the following information:

• The coordinates

• The auxiliary continuous variable name(s) (defined in 'auxnames')

• The auxiliary categorical variable name(s) (defined in 'auxnames')

The number of coordinates, as well as the number of auxiliary variables must match between 'dbin' and 'dbout'.

Value

The target Db where the resulting variable has been added.

---