RapidMiner Studio provides the Log
operator, which we have already seen being used in the previous chapters. Of all the operators, this is the one that I use a great deal, both for debugging and for creation of data.
Dealing with logging first, the Log
operator can be inserted anywhere in a process and is configured to output the parameters or values associated with another operator somewhere in the process. For example, the screenshot that follows shows some example parameters for the Log
operator:
The left-most column becomes the column name in the log, the second column is the name of the operator within the process, the third column is the type of information (either value or parameter), and the final column is the name of the information to log, which is filled in automatically with valid options by the RapidMiner Studio GUI. The value option is used to log the result of the execution of an operator, whereas the parameter option is used to log the parameters to the operator that controls its working.
The values of macros can be output using the Provide Macro as Log Value
operator and this operator is then referred to by its name in the Log
operator. An example of the output produced from the preceding screenshot is shown in the next table. The operator called Loopy is providing the value of a macro and this is accessed in the Log
operator by the selection of the macro value. The other entries are all derived from the parameters for various operators.
The value of a specific example of an attribute within an example set can be output using the Extract Log Value
operator and as with the Provide Macros as Log Value
operator, this is then accessed in the Log
operator.
The resulting log can be converted to an example set if a more permanent record needs to be kept for further analysis or if the example set needs to be used later as part of the normal processing being done by the process.. This is done using the Log
to Data
operator.