Displaying 1 result from an estimated 1 matches for "cnct_o".
2012 Sep 11
0
Question about logistic regression with ordered factor variable using the rms package (prev.Design)
...¦logit¡¦) ¡¨.
This works fine when test .value and concentration (cnct) are continuous
vairables.
However, concentration is in fact a grouping variable over 5 experiments
with 5 concentrations
( 25, 50, 100, 200 & 400).
Therefore I believe concentration to be an ordered factor ( in model :
cnct_o).
To make this model I used the ¡§rms¡¨ library (previously known as Design)
and functions lrm (or Glm).
The lrm (or Glm) returns the odds for disease, the ¡§inv.logit (odds) ¡¨
gives the probability of disease, but I have to do this with the Predict
function of the ¡§rms¡¨ package.
#####
The r...