Classifying from Counts - Naive Bayes from Counts

Naive Bayes starts with integer class counts, not fitted coefficients. This lesson fixes the class totals that every later ratio must reuse.

highlighted = computed this step

Counts before formulas

The shown data have class A count 4 and class B count 2. The total row count is 6.

\operatorname{count}(A)=4,\quad \operatorname{count}(B)=2

This surface starts from integer class counts. Later fractions are ratios formed from these displayed integers.

\text{integer counts become exact ratios}

No feature evidence is used yet. The first boundary is the displayed count table itself.

\text{class counts fixed}