Loading [a11y]/accessibility-menu.js
A performance evaluation of data streams sampling algorithms over a sliding window | IEEE Conference Publication | IEEE Xplore

A performance evaluation of data streams sampling algorithms over a sliding window


Abstract:

Over the last few years, a large number of applications have appeared generating data streams for which the exhaustive storage require high costs. It was, therefore, nece...Show More

Abstract:

Over the last few years, a large number of applications have appeared generating data streams for which the exhaustive storage require high costs. It was, therefore, necessary to process these data in real-time without storing them. Real-time processing requires specifying the analysis tasks before the arrival of the data. Consequently, any future need will not be answered because all the received data are lost. That is why it is necessary to keep a data stream summary in order to provide an approximate answer to the future queries. This can be done using the sampling algorithms. This paper is a continuation of our previous work in which we studied the Chain sampling algorithm. In this paper, we discuss two other sampling techniques: Deterministic sampling and Simple Random sampling (SRS) and we compare their performance against that of Chain sampling. The results show that Chain sampling gives better results than SRS and Deterministic sampling in terms of execution time and sampling accuracy respectively.
Date of Conference: 18-20 April 2018
Date Added to IEEE Xplore: 04 June 2018
ISBN Information:
Conference Location: Jounieh, Lebanon

References

References is not available for this document.