Ordinal Regression Link Function CAUCHIT CLOGLOG LOGIT NLOGLOG

순서 회귀분석 Ordinal Regression

Link Function CAUCHIT CLOGLOG LOGIT NLOGLOG PROBIT Cauchit function f(x) = tan(π(x – 0. 5)). latent variable has many extreme values Complementary log-log function f(x) = log(– log(1 – x)). higher categories more probable Logit function f(x) = log(x / (1 – x)). This is the default link function. evenly distributed categories Negative log-log function f(x) = –log(– log(x)). lower categories more probable Probit function f(x) = Φ -1(x), where Φ -1 is the inverse standard normal cumulative distribution function. latent variable is normally distributed

모형 적합도(전체) b 0 ui =(1, X 1 i, X 2 i, …Xk, i ) b 1 E(ui) = Yi + ei . . bk 확률밀도 함수 f(X: β 1, β 2, … βk) 실현 값 X 1, X 2 … Xm Likelihood Function When sample size m is large H 0 : β= 0 Пf(Xi; β 1, β 2, … βk) = L(β 1, β 2, … βk) or