A Simple, Yet Effective and Efficient, Sliding Window Sampling Algorithm

Xuesong Lu, Wee Hyong Tok, Chedy Raïssi, Stéphane Bressan

Abstract: Sampling streams of continuous data with limited memory, or reservoir sampling, is a utility algorithm. Standard reservoir sampling maintains a random sample of the entire stream as it has arrived so far. This restriction does not meet the requirement of many applications that need to give preference to recent data. The simplest algorithm for maintaining a random sample of a sliding window reproduces periodically the same sample design. This is also undesirable for many applications. Other existing algorithms are using variable size memory, variable size samples or maintain biased samples and allow expired data in the sample.