Create list of parameters for cross validate NEMoE

This function create parameters that put into cvNEMoE function.

createCVList(
  g1 = 10,
  lambda2_seq = c(0.005, 0.014, 0.016, 0.023, 0.025),
  shrink = 0.5,
  crit_eval = "all",
  crit_sel = "auc",
  track = FALSE,
  itmax_lambda = 3,
  itmax_cv = 20,
  itmax_fit = 50
)

Arguments

g1

Numbers of parameters lambda1 for validation.
lambda2_seq

A vector of candidates of lambda2.
shrink

A number of shrinkage of selected lambda1. By default is 0.5.
crit_eval

A vector of method for Evaluation metric of NEMoE object. Can be chosen from statistics "AIC", "BIC", "ICL1", "ICL2", "eBIC", "mAIC", "mBIC", "mICL1", "mICL2" and cross validation result "accuracy", "D.square", "TPR", "TNR", "F1" and "auc". If method = "all", all of the evaluation metric will be use. By default is "all.
crit_sel

Method use for select parameters. Should be a member in crit_eval.
track

Whether output the result of choosing parameters lambda1.
itmax_lambda

Maximal iterations of generation lambda1. By default is 3.
itmax_cv

Maximal iterations of calculate cross validation metric. By default is 20.
itmax_fit

Maximal iterations of fitting NEMoE inside the evaluation. By default is 50.

Value

A list contain parameters in fitting NEMoE.

See also

NEMoE_buildFromList, NEMoE_buildFromPhyloseq, fitNEMoE, calcCriterion, cvNEMoE

Examples

cvparams = createCVList(g1 = 10, lambda2_seq = c(0.005, 0.015, 0.02))