RSNNS (version 0.4-5)

SnnsRObjectMethodCaller: Method caller for SnnsR objects

Description

Enable calling of C++ functions as methods of SnnsR-class objects.

Arguments

x

object of class SnnsR-class

name

function to call

Details

This function makes methods of SnnsR__ and SnnsCLib__ accessible via "$". If no SnnsR__ method is present, then the according SnnsCLib__ method is called. This enables a very flexible method handling. To mask a method from SnnsCLib, e.g. to do some parameter checking or postprocessing, only a method with the same name, but beginning with SnnsR__ has to be present in R. See e.g. SnnsRObject$initializeNet for such an implementation.

Error handling is also done within the method caller. If the result of a function is a list with a member err, then SnnsCLib__error is called to use the SNNS kernel function to get the corresponding error message code and an R warning is thrown containing this message.

Furthermore, a serialization mechanism is implemented which all models present in the package use to be able to be saved and loaded by R's normal save/load mechanism (as RData files).

The completely trained object can be serialized with

s <- snnsObject$serializeNet("RSNNS_untitled")

snnsObject@variables$serialization <- s$serialization

For the models implemented, this is done in SnnsRObject$train. If the S4 object is then saved and loaded, the calling mechanism will notice on the next use of a function that the pointer to the C++ SnnsCLib object is nil, and if a serialization is present, the object is restored from this serialization before the method is called.