Displaying 1 result from an estimated 1 matches for "afactorimpacts".
2007 May 10
1
Follow-up about ordinal logit with mixtures: how about 'continuation ratio' strategy?
...# linear predictor
### eta = 0.3 * x + 0 (afactor==1)+ 0.4 * (afactor==2) + 0.5
(afactor==3)+ 0.9 (afactor==4)
### and the "cutpoints" between categories are -2.3, -0.5, and 0.66.
N <- 1000
A <- -1
B <- 0.3
cutpoint <- c( -2.3, -0.5, 0.66 )
afactor <- gl (n=4, k = N/4)
afactorImpacts <- c( 0, 0.4, .5, .9 )
### Make "first" impact 0 so it does not confuse our study of
intercept & cutpoints.
### Any other value gets up getting added into the intercept and/or
first cutpoint.
set.seed(333)
x <- 1 + 10 * rnorm(N)
eta <- B * x + afactorImpacts[as.numeric(afa...