I am seeking to evaluate logistic regression against a rule-based (nested if-else statements) classification in Orange.
The steps I've taken to investigate this are as follows:
- I could not identify any widgets that would implement this. The closest I could find was the interactive tree from version 2.7, but this widget seems to have been discontinued.
- I've attempted to generate the results I'm looking for by editing the in_data object produced by creating a tree, sending the results to the "Test and Score" widget, and then editing the results in "Python Script" to create an out_data object. However, the out_data object I produce won't work within the confusion matrix.
Is there an easier way to implement this rule-based criterion in Orange? If not, are there some other steps I can use to implement it within the Python script?