overfitting


정의: In mathematical modeling, overfitting is the production of an analysis that corresponds too closely or exactly to a particular set of data, and may therefore fail to fit to additional data or predict future observations reliably. An overfitted model is a mathematical model that contains more parameters than can be justified by the data. In the special case of a model that consists of a polynomial function, these parameters represent the degree of a polynomial. The essence of overfitting is to unknowingly extract some of the residual variation as if that variation represents the underlying model structure.


📄 키워드 상세정보

핵심 연구 분야Strategies
주요 연도2024년
주요 연관 키워드challenge
좋아요 수0

키워드별 논문 목록 (11건)