The possibility to calculate the median of a stream of data could be useful for different analytics like the estimation of the detector background. Unlike the mean one cannot calcumate the exact mean of a dataset in a single pass, however it is possible to extimate the median (or any other quantile).
A good candidate of an efficient single-pass algorithm to approximate the mean could be the $P^2$-algorithmen proposed by Jain and Chlamtac.
This algorithm has several useful features:
it produces estimates dynamically from incoming data
thus, no knowledge of the minimum and maximum values is required
the storage requirements are fixed, small and independent from the size of the sample size
the estimates perform well for sample sizes bigger than 100 samples in most cases
but no garantee on the quality of the observation is given and the algorithm does not perform well when there are discontinuities close to the approximated quantile
The possibility to calculate the median of a stream of data could be useful for different analytics like the estimation of the detector background. Unlike the mean one cannot calcumate the exact mean of a dataset in a single pass, however it is possible to extimate the median (or any other quantile). A good candidate of an efficient single-pass algorithm to approximate the mean could be the $P^2$-algorithmen proposed by Jain and Chlamtac.
This algorithm has several useful features: