spam: Estimates sparse (mulitivariate) additive model

In dajmcdon/spasm: Implements the Laplace-Gaussian filter for sparse, additive, state-space models

Description

Estimates sparse (mulitivariate) additive model

Usage

spam(Y, X, df = 10, which.lam = 10, returnall = FALSE, ...)

Arguments

Y	response, Nxd matrix
X	predictors, Nxp matrix
df	the number of spline basis functions to use, defaluts to 10
which.lam	component of the tuning parameter vector for saving coefficients
returnall	if TRUE, returns extra information, see below
...	optional arguments to samQL

Value

a list with components

knots

a list with components interior, a (df-1)xp matrix and boundary, a 2xd matrix

B.hat

a (d)x(1+df*p) matrix containing the retained coefficients (intercept in the first column)

range.info

a list with components min and range of size px1 each describing the predictors

spam.out

if returnall is TRUE, a list of length d containing all output from samQL. Useful for the predict or fitted methods

X

the original X input

Examples

someData = generateSPASM(100,3,4)
SPAMfit = spam(someData$y, someData$x)

dajmcdon/spasm documentation built on May 6, 2019, 1:31 a.m.