recover_data: Perform feature selection on pointers and import data back in...
In linogaliana/OpenCancer: Build statistical models to understand cancer causes

Description Usage Arguments

recover_data has been designed to easily perform feature selection by LASSO on pointers and import selected features as well as additional relevant variables back in the memory

1
2
3

recover_data(X, yvar = "incidence", labelvar = c("cancer", "age",
  "Country_Transco", "year", "area.x", "area.y"), crossvalidation = T,
  nfolds = 10, returnplot = F, ncores = 1)

`X`	A `big.matrix` object
`yvar`	Name of the variable we want to explain in `X`
`labelvar`	Variable names of columns excluded from the set of covariates in LASSO. All variables included in `labelvar` will be added to the set of selected features imported back in memory
`crossvalidation`	Should we perform a cross-validated LASSO ? `TRUE` or `FALSE`
`nfolds`	Number of folds for cross-validation. Ignored if `crossvalidation = FALSE`. If `nfolds` equals the number of observations, leave-one-out cross validation is performed
`returnplot`	Should we return a plot of the relationship between quadratic risk and $\lambda$ parameter? `TRUE` or `FALSE`
`ncores`	Number of cores used for computations. If `ncores>1`, parallel processing is used