Good day everyone!
Can you please help me understand the prediction results from binary classifications?
I used a dataset with 5 different methods. Simple model with steps in SPSS Modeler: Load data -> Type -> Partition -> model nodes -> Analysis
When model nuggets generated, I use them into a validation data set.
The validation results generated 5 separate Excel files (one for each prediction method).
My question is about the 2 columns (F and G) added:
- Column F seems like the prediction result
- Column G seems like the likelihood of the prediction
If so, I can understand in Discriminant, the highest likelihood corresponds to the class "1". Same in Logistic.
However, in Neural net, prediction class "1" doesn't have the highest likelihood. Neither does SVM. And in Bayes, the likelihood 1 was predicted as class "0".
What's the right way to understand the Column F and Column G?
Thank you!
@DAVID WEST,
@Eitel Lauría,
@Nicolas Rafael Ascanio Pena,
@Stephen Lambert, can you please help me here? Thanks!
#SPSSModeler