METHOD COMPUTER DEVICE AND COMPUTER PROGRAM FOR ENSEMBLE PREDICTION TO IMPLEMENT AUTOMATED MACHINE LEARNING

Disclosed are a method, a computer device, and a computer program for ensemble prediction to implement automated machine learning. The method for ensemble prediction comprises the steps of: inferring a representation vector by vectorizing given input data; assigning a weight to each basic model base...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: HONG YOUNGTEAC, LEE JUHO, PARK YOUNG JIN, KIM KYUNG MIN, KIM JIHOON
Format: Patent
Sprache:eng ; kor
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Disclosed are a method, a computer device, and a computer program for ensemble prediction to implement automated machine learning. The method for ensemble prediction comprises the steps of: inferring a representation vector by vectorizing given input data; assigning a weight to each basic model based on the expression vector, targeting a plurality of basic models which predict the future for the input data; and determining model combination through ensemble prediction based on the weight for each of the basic models. The present invention can adaptively select and assign the basic model. 자동기계학습을 구현하기 위한 앙상블 예측 방법, 컴퓨터 장치, 및 컴퓨터 프로그램이 개시된다. 앙상블 예측 방법은, 주어진 입력 데이터를 벡터화함으로써 표현 벡터를 추론하는 단계; 상기 입력 데이터에 대한 미래를 예측하는 복수 개의 기본 모델을 대상으로 상기 표현 벡터를 기초로 각 기본 모델의 가중치를 부여하는 단계; 및 상기 기본 모델 별 가중치에 기초한 앙상블 예측을 통해 모델 조합을 결정하는 단계를 포함할 수 있다.