set_class_box_param — Set system parameters for classification.
set_class_box_param is obsolete and is only provided for
reasons of backward compatibility. New applications should use the
MLP, SVM, KNN or GMM operators instead.
set_class_box_param modifies parameter which manipulate the training
sequence while calling
learn_class_box. Only parameters of the
classifier are modified, all other classifiers remain unmodified.
'min_samples_for_split' is the number of examples at least which have to
train in one cuboid of this classifier, before the cuboid is
allowed to divide itself.
'split_error' indicates the critical error. By its exceeding the cuboid
divides itself, if there are more than 'min_samples_for_split' examples to
'prop_constant' manipulates the extension of the cuboids. It is proportional
to the average distance of the training examples in this cuboid to the
center of the cuboid.
extension x prop = average distance of the expectation value. This relation is valid in every dimension. Hence inside a cuboid the dimensions of the feature space are supposed to be independent.
The parameters are set with problem independent default values, which
must not modified without any reason. Parameters are only important during a
learning sequence. They do not influence on the behavior of
'min_samples_for_split' = 80,
'split_error' = 0.1,
'prop_constant' = 0.25.
This operator modifies the state of the following input parameter:
During execution of this operator, access to the value of this parameter must be synchronized if it is used across multiple threads.
ClassifHandle(input_control, state is modified) class_box
Handle of the classifier.
Name of the wanted parameter.
Default value: 'split_error'
Suggested values: 'min_samples_for_split', 'split_error', 'prop_constant'
→(real / integer)
Value of the parameter.
Default value: 0.1
read_sampset returns 2 (H_MSG_TRUE).