My plan is to use a published nomogram to predict events in my data set. The question is, how do I derive logistic coefficients and intercept from the nomogram?
5
-
$\begingroup$ The nomogram you want to use has a reference somewhere? It's possible the coefficients of the logistic regression are clearly indicated in the reference work $\endgroup$– GGACommented Jan 24, 2016 at 13:51
-
$\begingroup$ No. And I have asked them to provide the coefficients but to no avail. $\endgroup$– ayolCommented Jan 24, 2016 at 13:54
-
$\begingroup$ Can you post here the reference? A spurious approach could be constructing a dataset calculating the results and infer a regression equation in that dataset $\endgroup$– GGACommented Jan 24, 2016 at 13:55
-
$\begingroup$ Take this for example. Not the one that I want to use but close enough. sciencedirect.com/science/article/pii/S0167814011003343 $\endgroup$– ayolCommented Jan 24, 2016 at 14:04
-
$\begingroup$ I think you could build a data set with different values and compute a regression equation between them, maybe a digital approach using the lenght and absolute position of the Lines calculated using the nomogram image would be more precise $\endgroup$– GGACommented Jan 24, 2016 at 14:10
Add a comment
|