How to ensure the same probability for samples of unknown total number

1. Randomly extract a number from an unknown large number

Solution: If this number is choice, replace it with the probability of 1/I for the number of I

Choice = 1

Replace choice = 2 with a probability of 1/2

Replace choice = 3 with a probability of 1/3

...

Until data traversal is complete

2. Extract K from an unknown large number

Solution: add the number of I to the sampling set with the probability of K/I. If the sampling set is full, replace one randomly.

If you want to extract 10

Then 1, 2, 3 ...... 10 is the initialization set

For 11, join the set with a probability of 10/11 and replace one randomly

For 12, join the set with a probability of 10/12 and replace one randomly

...

Until data traversal is complete

 

 

Http://archive.cnblogs.com/a/1949762/