Ranking Events Based on a Random Sample
摘要
When a random sample of items is classified into categories, which are then ranked from the most frequent to the least, a decision made too early is prone to being wrong. How large must a sample be so that we are reasonably assured of a correct ranking?