elastic: Elastic-net modeling of ER objects.

In ER: Effect + Residual Modelling

Elastic-net modeling of ER objects.

Description

Elastic-net modeling of ER objects.

Usage

elastic(er, ...)

## S3 method for class 'ER'
elastic(
  er,
  effect,
  alpha = 0.5,
  newdata = NULL,
  validation,
  segments = NULL,
  measure = measure,
  family = family,
  ...
)

Arguments

er	Object of class ER.
...	Additional arguments for pls::cvsegments.
effect	The effect to be used as response.
alpha	The elasticnet mixing parameter.
newdata	Optional new data matrix for prediction.
validation	Optional validation parameters.
segments	number of segments or list of segments (optional)
measure	Type of performance summary, default = 'class' (see glmnet)
family	Type of model response, default = 'multinomial'.

See Also

ER, pls and confints.

Examples

## Multiple Sclerosis data
data(MS, package = "ER")
er <- ER(proteins ~ MS * cluster, data = MS)
elasticMod <- elastic(er, 'MS', validation = "CV")
sum(elasticMod$classes == MS$MS)
plot(elasticMod)            # Model fit
plot(elasticMod$glmnet.fit) # Coefficient trajectories

# Select all proteins with non-zeros coefficients
coefs     <- coef(elasticMod,s='lambda.min',exact=TRUE)
(selected <- rownames(coefs[[1]])[unique(unlist(lapply(coefs,
                      function(x)which(as.vector(x) != 0))))][-1])


## Diabetes data
data(Diabetes, package = "ER")
er.Dia <- ER(transcriptome ~ surgery * T2D, data = Diabetes)
elasticMod <- elastic(er.Dia, 'T2D', validation = "LOO")

ER documentation built on Oct. 11, 2022, 1:07 a.m.