TClassesListItem Class

TPersistentTClassesListItem

type TClassesListItem = class(TPersistent);

Abstract class for items of a list of clasification classes.

This is an abstract class defined to manage data for a classification class. Each classification class holds a list of attributes (parameters) and the related statistics. The statistics is obtained during the learn run. This statistics is then used when comparing each class to the new (to be classified) example and computing the response of each class. The classification class with the strongest response wins.

Methods

NameDescription
ClassifyClassify record with mixed real and discrete attributes.
ClassifyDiscreteClassify record with discrete attributes.
ClassifyFloatClassify record with real valued attributes.
CreateCreate the component and pass a TClassesList object as an owner.
LearnLearn a new record containing both discrete and real attributes.
LearnDiscreteLearn the values of discrete attributes of the new example.
LearnFloatLearn the values of continuous (real valued) attributes of the new example.
LoadFromStreamLoad the object from the stream.
PriorProbabilityCompute prior probability after the learning has been completed.
ResetReset learned data.
ResetWeightsReset the Weight field of every attribute back to one.
SaveToStreamSave the object to stream.
SoftResetResets learned data without reseting weights and excluded attributes.

Constants

NameTypeDescription
DiscreteATDiscreteRecordArrayStores the statistics about the discrete attributes.
FloatATFloatRecordArrayStores the statistics about the continuous (real valued) attributes.