update.trainOcc: Update or Re-fit a Model
In benmack/oneClass: One-class classification in the absence of test data

Description Usage Arguments Details Value Examples

Over-write the tuning parameter selection process. You can also calculate new performance metrics given the hold-out predictions are stored in object.

## S3 method for class 'trainOcc'
update(object, modParam = NULL, modRow = NULL,
  modRank = NULL, metric = NULL, puSummaryFunction = NULL,
  aggregatePredictions = FALSE, newMetricsOnly = FALSE,
  allowParallel = TRUE, verbose = FALSE, ...)

`object`	a object of class `trainOcc`.
`modParam`	a data frame with the desired model parameters, or
`modRow`	the index of the model, i.e. the row in the `object$results` table, or
`modRank`	the model at the modRank-th position after sorting metric a performance metric (can be specified via argumetn `metric`).
`metric`	The metric used to select the model, e.g. in conjunction with `modRank` or new summary functions. if `NULL` the metric in `object$metric` is used.
`puSummaryFunction`	a function that can be passed to the `summaryFunction` argument of `trainControl`. This updates the `trainOcc$results` data frame with the new metric(s) returned from the function. (Only be done when hold-out predictions are stored in `object`.)
`aggregatePredictions`	The default is `FALSE`, which means that the performance metric is calculated for each set of hold-out predictions and the metrics are then aggregated. The mean and the standard deviation is returned in the `object$results` table. If `TRUE` the hold-out predictions are first aggregated and the `puSummaryFunction` is calculated once. Of course, no standard deviation of the performance metric can be calculated. Metrics calculated this way get the suffix `AP` in the `object$results` table.
`newMetricsOnly`	logical with default set to `FALSE`. Set to `TRUE` if the metrics already contained in the results table should be removed.
`allowParallel`	if true and if a parallel backend is registered the calcualtions are run parallel
`verbose`	A logical for printing information about the updated model.
`...`	other arguments that can be passed to update.train

update.train

an updated trainOcc object.

## Not run: 
data(bananas)

tocc <- trainOcc(x=bananas$tr[, -1], y=bananas$tr[, 1], method="ocsvm")

## update to the highest ranked model according to metric puAuc
tocc <- update(tocc, modRank=1, metric="puAuc")

## End(Not run)