Estimating a number of unique values in a list

摘要

A method determines a number of unique values in a sample of a list of values and estimates a number of the unique values for an unsampled portion of the list of values. The method estimates a number of the unique values in the list by adding the number of unique values in the sample to the number of the unique values in the unsampled portion.

主权项

1. A method executed by a computer system, comprising:
storing, in the computer system, a list of values; determining, with the computer system, a number of unique values and a frequency of the unique values for a sample of the list of values; estimating, with the computer system, a number of unique values and a frequency of the unique values for an unsampled portion of the list of values, by adding a value based on a distribution family for the frequency of the unique values for the sample to a value based on a conditional probability distribution family for the frequency of the unique values in the unsampled portion given the frequency of the unique values for the sample; and estimating, with the computer system, a number of unique values in the list by adding the determined number of unique values in the sample to the estimated number of unique values in the unsampled portion.