Open AccessSubsampling Method for Big Data in Poisson Regression
Author(s) -
A. N. I. Pradana,
A’yunin Sofro
Publication year - 2019
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1417/1/012013
Subject(s) - akaike information criterion , poisson regression , poisson distribution , statistics , computer science , sample (material) , sample size determination , regression analysis , regression , big data , sampling (signal processing) , data mining , mathematics , population , chemistry , demography , filter (signal processing) , chromatography , sociology , computer vision
Big Data is very large, too fast and complex that cannot be managed with traditional tools; and must use new methods and tools to process it. Subsampling method is one method that can be used to solve big data problems. This paper focuses on estimation in Poisson Regression using Maximum Likelihood Estimation (MLE). In this study, it will be discussed about the implementation of the Subsampling method of The Demand for Medical Care data. To understand the performance of subsampling method, the paper provided a scenario with different sample sizes. The results of this study obtained parameter estimates and Poisson regression models on each subsample. Based on the Akaike Information Criterion, the best model is providing the smallest value. From the scenario, the sample size of the subsampling which is close to sampling tend to have smaller value of AIC.