다음을 통해 공유


GeneralityLossRule(Single) 생성자

정의

일반성이 손실된 경우 학습 프로세스를 종료할 수 있는 규칙을 만듭니다. 일반성이 손실되면 유효성 검사에서 지정된 점수가 증가하기 시작합니다.

public GeneralityLossRule(float threshold = 0.01);
new Microsoft.ML.Trainers.FastTree.GeneralityLossRule : single -> Microsoft.ML.Trainers.FastTree.GeneralityLossRule
Public Sub New (Optional threshold As Single = 0.01)

매개 변수

threshold
Single

(현재) 유효성 검사 점수와 최적 기록 값 사이의 최대 간격(예: 1%의 경우 0.01, 50%의 경우 0.5)입니다.

적용 대상