Setting of system parameters for classification.
set_classif2_param modifies parameter which manipulate the training sequence while calling learn_classif2. Only parameters of the current classificator are modified, all other classificators remain unmodified. 'min_samples_for_split' is the number of examples of trainings which have to result at least in one cuboid of this classificator, before the cuboid is allowed to divide itself. 'split_error' indicates the critical error. If it is exceeded the cuboid divides itself, if it encountered already more than 'min_samples_for_split' training examples. '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. More detailed:
extension x prop = average distance of the anticipation value.This relation is valid in every dimension. Hence inside a cuboid the dimensions of the attribute area are supposed to be independent.
The parameters are set with widely problem independent default values, which must not modified without any reason. Parameters are important during a learning sequence only. They exert no influence on the behavior of enquire_classif2.
Default setting: 'min_samples_for_split' = 80,
'split_error' = 0.1,
'prop_constant' = 0.25
Flag (input_control) |
string -> string |
Name of the wanted parameter. | |
Default value: 'split_error' | |
Suggested values: 'min_samples_for_split', 'split_error', 'prop_constant' |
Value (input_control) |
number -> real / integer |
Value of the parameter. | |
Default value: 0.1 |
read_sampset returns TRUE. An exception handling is raised, if there is no current classificator (see set_classif2) .
create_classif2, enquire_classif2, get_classif2, set_classif2
learn_classif2, test_sampset2, write_classif2, close_classif2, free_sampset
enquire_classif2, get_classif2_param, learn_classif2, set_classif2