Hey @vineetchanana, no when we talk about P(A|B) then it mean probability of occurrence of A when B has already occurred. So when we calculate P(xi|y=c) then it means, y=c has already occurred, means we have extracted all example where class or y = c. So after this we only need to calculate probability of xi word. Since probability is always calculated using formula number of occurrence divided by total number of items. So here for denominator, total number of items is the dictionary/vocab size from all examples having class = c
Hope this resolved your doubt.