Class TCalcSpvMLP

Unit

UCompSpvMLPerceptron

Declaration

type TCalcSpvMLP = class(TCalcSpvLearning)

Description

classe du calculateur de MLP

Hierarchy

TCalcSpvLearning

Fields

	Name	Description
	`FInputLayerSize`	taille de l'entrée, càd de la première couche - ici le nombre de descripteurs
	`FNbNeuronsPerLayer`	nombre de neurones par couche
	`FStatsDescriptors`	statistiques sur les variables

Methods

Overview

	`constructor create(prmOpSpv: TOpPrmSpvLearning; prmClass: TAttribute; prmDescriptors: TLstAttributes; prmAllAttributes: TLstAttributes); override;`
	`function afterLearning(examples: TExamples): boolean; override;`
	`procedure afterProcessPatterns(); virtual;`
	`function beforeLearning(examples: TExamples): boolean; override;`
	`procedure beforeProcessPatterns(); virtual;`
	`procedure classification(example: integer; var response: TTypeDiscrete; var pDist: TTabFrequence); override;`
	`function coreLearning(examples: TExamples): boolean; override;`
	`procedure createStructures(); override;`
	`procedure destroyStructures(); override;`
	`function getHTMLResults(): string; override;`
	`procedure InputPattern(example: integer); virtual;`
	`procedure SetAttToBlank(example: integer; prmAttEval: integer); virtual;`

Description

constructor create(prmOpSpv: TOpPrmSpvLearning; prmClass: TAttribute; prmDescriptors: TLstAttributes; prmAllAttributes: TLstAttributes); override;

function afterLearning(examples: TExamples): boolean; override;