I' ve been reading some scientific works and I don't understand how nomograms are constructed from logistic regression models.
In the article:
Development and Validation of an Early Scoring System for Prediction of Disease Severity in COVID-19 Using Complete Blood Count Parameters
I also provide link to download from my Google Drive -> https://drive.google.com/file/d/1xEoYh6RZJ3bsqdaCbfE943E11xFtgmeP/view?usp=sharing
In the article they work with COVID19 dataset describing two target values. Patient died from covid or survived. Along with preprocessing phase and feature selection as well as statistical characteristics they choose 10 blood parameters on which the algorithms will be trained and evaluated. [well not only blood as age is here too]
After training it turns out that Logistic Regression model has the best results, so it is chosen for nomogram creation.
The point I can't undestand is how they come up with Death probability and this:
I get all the part conserning preprocessing, feature selection and statistical characterictics.