Construct confusion matrix for multi-class multi-column classification
I would like to build a confusion matrix for multi-class multi-column classification for calculating precision, recall and F1. One idea is to build it from all the combinations that go into the training and testing kit, for example.
A1 A2A3 A1A3
A1 x x x
A2A3 x x x
A1A3 x x x
Another idea is to plot it like for a simple label classification, but use double values for matrix values, for example.
A1 A2 A3
A1 double double double
A2 double double double
A3 double double double
The question in this case is how to calculate these values meaningful?
Does anyone have any experience with creating such matrices? Which version is more rational? If there is any other way to construct such a confusion matrix, it would be nice to hear from you.
Greetings, Andrey
source to share
If anyone is also interested in this, this is how it works for me: I used the first idea and calculated the measures based on the labels due to the description from: Gj. Madjarov et al., An Extensive Experimental Comparison of Multilevel Learning Methods, The Recognition Pattern (2012).
The relevant code can be found in dkpro-tc (DKPro Text Classification Framework) in the evaluation unit.
source to share