search for




 

A study on pattern estimation of repeated measured packets of three mobile portals
Journal of the Korean Data & Information Science Society 2018;29:251-65
Published online January 31, 2018
© 2018 Korean Data and Information Science Society.

Gui-Yeol Ryu1

1Department of Computer Science, SeoKyeong University
Correspondence to: Professor, Department of Computer Science, SeoKyeong University, Seoul 136-704, Korea. E-mail: gyryu@skuniv.ac.kr
Received October 11, 2017; Revised November 2, 2017; Accepted November 2, 2017.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
The purpose of this paper is to build a proper model for packet amount of mobile portals such as Naver, Daum, and Nate. We considered three time series models: the first one is of raw data, the second one is of multiplicative decomposition data, and the last one is of additive decomposition data. The time series models have been built using autoregressive model, in which predictors were repetition, date, week, and month. We collected 3,246 cases by measuring the sixth per access from September, 2012 to June, 2017. The significant predictors and AR parameters have been selected respectively by backward elimination and by significant test. Comparing the three models by AIC and MSE, we selected the multiplicative decomposition model for Naver and additive decomposition model for Daum, raw data model for Nate. The models are proper from Box-Ljung statistic and residual ACF and PACF. Total R2 are 78.83% for Naver, 75.53% for Daum, and 80.55% for Nate. The results of this paper show that it is necessary to improve the accuracy of estimation by comparing and analyzing repeatedly measured time series data by various methods such as seasonal decomposition model.
Keywords : Additive seasonal decomposition, autoregressive model, mobile portal, multiplicative seasonal decomposition, packet amount.