Reg.ispline {zetadiv}

Perform an I-spline regression

Description

Evaluates the I-splines for all variables of a data frame of predictor variables, and perform a generalised linear regression with constraint on the parameters.

Usage

Reg.ispline(
  response,
  predictor,
  order.ispline = 2,
  kn.ispline = 1,
  family = stats::gaussian(),
  method.glm = "glm.fit.cons",
  cons = 1,
  cons.inter = 1,
  control = list(),
  Plot = TRUE,
  lty = NULL,
  lwd = 1
)

Arguments

response

A vector of numeric values representing the response variable.

predictor

A data frame of numeric variables representing the predictors.

order.ispline

Order of the I-spline.

kn.ispline

Number of knots in the I-spline.

family

A description of the error distribution and link function to be used in the glm model (see family for details of family functions).

method.glm

Method used in fitting the generalised linear model. The default method
"glm.fit.cons" is an adaptation of method glm.fit2 from package glm2 using a constrained least squares regression in the reweighted least squares. Another option is "glm.fit2", which calls glm.fit2; see help documentation for glm.fit2 in package glm2.

cons

type of constraint in the glm if method.glm = "glm.fit.cons". Default is 1 for positive coefficients on the predictors. The other option is -1 for negative coefficients on the predictors.

cons.inter

type of constraint for the intercept. Default is 1 for positive intercept, suitable for Gaussian family. The other option is -1 for negative intercept, suitable for binomial family.

control

As for glm.

Plot

Boolean value indicating if the I-splines must be plotted.

lty

Line types to be used in the plotting. If nothing is provided, lty is a sequence of integers from 1 to the number of variables used for the computation of msgdm.

lwd

Line width.

Details

Reg.ispline performs a non-linear regression using a combination of GLM and I-splines. It can, for example, be used to compare regression outputs when using MS-GDM with I-splines on environmental variables and biotic variables as in Zetya.msgdm to the same regression approach without environmental variables.

Value

Reg.ispline returns a list of the following elements:

splines

A data frame in which each columns contains the value resulting from the transformation of the predictors into individual I-splines. The number of columns of splines is the number of predictors times the number of splines (determined as the sum of order.ispline and kn.ispline).

spline

A data frame in which each columns contains the value resulting from the combinations of the individual I-splines. This combination is obtained by multiplying the coefficients of model and the values of the individual I-splines splines

.

model

A glm model using response as the response variable, and splines as the predictors

.

References

Ramsay, J. O. (1988). Monotone regression splines in action. Statistical Science, 425-441.

See Also

Zeta.msgdm,Ispline

Examples

utils::data(Marion.species)
xy.marion <- Marion.species[1:2]

data.spec.marion <- Marion.species[3:33]

##random other communities
data.spec.marion2a <- data.spec.marion
data.spec.marion2a[which(data.spec.marion2a==1,arr.ind=TRUE)] <- 0
for(i in 1:ncol(data.spec.marion2a))
  data.spec.marion2a[sample(nrow(data.spec.marion2a),8),i] <- 1
data.spec.marion2b <- data.spec.marion
data.spec.marion2b[which(data.spec.marion2b==1,arr.ind=TRUE)] <- 0
for(i in 1:ncol(data.spec.marion2b))
data.spec.marion2b[sample(nrow(data.spec.marion2b),8),i] <- 1

dat.spec.tot <- list(data.spec.marion,data.spec.marion2a,data.spec.marion2b)
zeta.tot <- Zeta.order.mc.mult(data.spec=dat.spec.tot,order=3,sam=200)
zeta.splines <- Ispline(zeta.tot$zeta.val[,2:3])
data.tot <- data.frame(zeta.val=zeta.tot$zeta.val[,1],zeta.splines$splines)

dev.new()
Reg.ispline(response = zeta.tot$zeta.val[,1], predictor = zeta.tot$zeta.val[,2:3], lwd=2, cons=1)