Extracting sample data based on poisson distribution
Authors:
- 1 July 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 2, 374-378
Abstract
Sampling methods are becoming in demand due to the rapid growth of big data applications. The term “Big Data” not only means the large size of data volume but also indicates the high speed of data generation, which plagues many existing data mining and analytic applications owing to the limited capability of processing large volume of data for real time analysis. Therefore, the demands for the use of sampling to generate summary data sets that support rapid queries are increasing according to Cormode and Duffield. The state-of-the art in sampling methods have been successfully applied to various areas including network traffic and social networks[1]. In this paper, a novel Poisson-based sampling method is introduced to provide a comprehensive data set for real time analysis. The proposed Poisson-based sampling method extends the previous Normal Distribution sampling method [2]. The experimental results show efficiency of the proposed method.Keywords
This publication has 7 references indexed in Scilit:
- Characterizations of discrete compound Poisson distributionsCommunications in Statistics - Theory and Methods, 2016
- Splitting Large Medical Data Sets Based on Normal Distribution in Cloud EnvironmentIEEE Transactions on Cloud Computing, 2015
- Sampling for big data2014
- Effect of separate sampling on classification accuracyBioinformatics, 2013
- Recent Techniques of Clustering of Time Series Data: A SurveyInternational Journal of Computer Applications, 2012
- The Google file systemACM SIGOPS Operating Systems Review, 2003
- On the Early History of the Singular Value DecompositionSIAM Review, 1993