java.lang.Object
weka.classifiers.evaluation.output.prediction.AbstractOutput
weka.classifiers.evaluation.output.prediction.XML
All Implemented Interfaces:
Serializable, OptionHandler

public class XML extends AbstractOutput
Outputs the predictions in XML.

The following DTD is used:

<!DOCTYPE predictions
[
<!ELEMENT predictions (prediction*)>
<!ATTLIST predictions version CDATA "3.5.8">
<!ATTLIST predictions name CDATA #REQUIRED>

<!ELEMENT prediction ((actual_label,predicted_label,error,(prediction|distribution),attributes?)|(actual_value,predicted_value,error,attributes?))>
<!ATTLIST prediction index CDATA #REQUIRED>

<!ELEMENT actual_label ANY>
<!ATTLIST actual_label index CDATA #REQUIRED>
<!ELEMENT predicted_label ANY>
<!ATTLIST predicted_label index CDATA #REQUIRED>
<!ELEMENT error ANY>
<!ELEMENT prediction ANY>
<!ELEMENT distribution (class_label+)>
<!ELEMENT class_label ANY>
<!ATTLIST class_label index CDATA #REQUIRED>
<!ATTLIST class_label predicted (yes|no) "no">
<!ELEMENT actual_value ANY>
<!ELEMENT predicted_value ANY>
<!ELEMENT attributes (attribute+)>
<!ELEMENT attribute ANY>
<!ATTLIST attribute index CDATA #REQUIRED>
<!ATTLIST attribute name CDATA #REQUIRED>
<!ATTLIST attribute type (numeric|date|nominal|string|relational) #REQUIRED>
]
>

Valid options are:

 -p <range>
  The range of attributes to print in addition to the classification.
  (default: none)
 -distribution
  Whether to turn on the output of the class distribution.
  Only for nominal class attributes.
  (default: off)
 -decimals <num>
  The number of digits after the decimal point.
  (default: 3)
 -file <path>
  The file to store the output in, instead of outputting it on stdout.
  Gets ignored if the supplied path is a directory.
  (default: .)
 -suppress
  In case the data gets stored in a file, then this flag can be used
  to suppress the regular output.
  (default: not suppressed)
Version:
$Revision: 8937 $
Author:
fracpete (fracpete at waikato dot ac dot nz)
See Also: