Hi, Should differing values of alpha have an impact on a binary, CNB classifier? Either they shouldn't or I'm doing something wrong. No matter what values I use, I get virtually identical results using the same training and test vectors. The confusion matrixes are identical. The only actual differences I've been able to find are negligible changes probably due to rounding in about 10% of the result vectors.
David -- David Engel [email protected]
