Description:
A, B, C and D arrive in order. Keep the first item; at arrival number n, replace the current item with probability 1/n. A survives with 1/2 x 2/3 x 3/4 = 1/4, and every other item also finishes with probability 1/4.
The sampling unit matters. This rule gives arriving items equal chances; an entity appearing several times still receives several opportunities.
An original illustrated concept lesson from OriginChainDB Console Academy; no live product result is claimed.
Reference: Vitter: Random Sampling with a Reservoir, section 2
https://db.cs.berkeley.edu/cs286/papers/reservoirsampling-toms1985.pdf
#OriginChainDB #ConsoleAcademy #DataEngineering #DeveloperEducation #DataLiteracy #SoftwareDevelopment #TechEducation #Shorts #ReservoirSampling #StreamSampling #Probability #RandomSampling #DataStreams #SamplingUnit #UniformSampling #Algorithms #RepeatedEvents #SampleSelection #Statistics #StreamingData
Share this link via
Or copy link