RDesc Portable

RDesc is an application software download free Windows that enables users to evade the usual annoyances encountered when seeking out files on hosting sites such as RapidShare, MegaUpload or MediaFire: banners and ads, queues, waiting times. With RDesc Portable, users can import a list of download links and the program will automatically manage their download and save them into a selected folder.

The resample class (makeResampleDesc()) encapsulates training and test sets generated by the given resampling strategy over a number of iterations. It also encapsulates optionally the grouping of resampling iterations. This is a convenient way of emulating the effect of repeated cross-validation (RCV) and to allow performance evaluation against previous iterations.

Typically, it is desirable that the proportion of classes in the training set remains the same as in the test set. Otherwise, observations of less frequent classes may be missing in some training sets and decrease the performance of the learner, or even lead to model crashes. To achieve this goal, it is possible to conduct stratified resampling (also known as "reweighted RCV"). To do so, set predict = "train" or predict = "both" in makeResampleDesc().

The function getRRPredictionList() returns a list of predictions split by data set (train or test) and resampling iteration. The return value is the mean prediction error over all resampling iterations, and can be used to compare models on similar data sets. In addition, a list of index vectors can be retrieved by calling getIndexVectors().