Evaluating NaiveBayes classifier by constructing confusion matrix

Question

I am trying to build a confusion matrix for blood Pressure patient where in actual number of patients predicted Blood pressure=0.18 and Not .82 Now the Naive Bayes classifier has Predicted no of patients having Blood pressure=0.410 and not 0.0913.I am new to this and i am not getting the idea of how to construct confusion matrix for the above problem.Please anyone could help me?

Answer 1

To compute the confusion matrix, you need to know the agreement on individual test samples.

You cannot compute it from the frequencies of the individual estimators alone.

Mathematically, you need the joint distribution, but you gave the edge distributions only.