[INU, 3-1] Data Science

Classification : 입력에 대해서 Regression : 입력에 대해 실수의 예측값을 예측하는 modelLogistic Regression : Regression 문제인데, 확률적인 요소가 들어간 model이번 chapter는 Classification에 대

A statistical method to study relationship between $x$ and $y$$x$ : covariate / predictor variable / independent variable / feature$y$ : response / de

Logistic Regression : 예측하는 값을 확률적으로 접근Classification 장점 + Regression 장점Linear Classification: Signal is thresholded at zero to produce +-1 outputFor b

Traditionally most studied model : feedforward neural networkComprises multiple layers of logistic regression modelsCentral idea주어진 입력으로부터 좋은 feature를

Function Signals == Feed ForwardPropagates forward Error Signals == Backward PropagationPropagates backwardForward PropagationBackward Propagation$d$