The Korean Journal of Applied Statistics (응용통계연구)

The Korean Statistical Society (한국통계학회)
권호 표지
DOI QR코드

DOI QR Code

Automatic order selection procedure for count time series models

계수형 시계열 모형을 위한 자동화 차수 선택 알고리즘

  • Ji, Yunmi (Department of Applied Statistics, Chung-Ang University) ;
  • Seong, Byeongchan (Department of Applied Statistics, Chung-Ang University)
  • 지윤미 (중앙대학교 응용통계학과) ;
  • 성병찬 (중앙대학교 응용통계학과)
  • Received : 2019.12.30
  • Accepted : 2020.01.12
  • Published : 2020.04.30
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, we study an algorithm that automatically determines the orders of past observations and conditional mean values that play an important role in count time series models. Based on the orders of the ARIMA model, the algorithm constitutes the order candidates group for time series generalized linear models and selects the final model based on information criterion among the combinations of the order candidates group. To evaluate the proposed algorithm, we perform small simulations and empirical analysis according to underlying models and time series as well as compare forecasting performances with the ARIMA model. The results of the comparison confirm that the time series generalized linear model offers better performance than the ARIMA model for the count time series analysis. In addition, the empirical analysis shows better performance in mid and long term forecasting than the ARIMA model.

본 논문은 시계열 일반화 선형 모형의 하나인 계수형 시계열 모형에서 중요한 역할을 하는 과거 관측값과 조건부 평균값의 차수를 자동으로 결정하는 알고리즘을 연구한다. 본 알고리즘은 ARIMA 모형의 차수를 기반으로 시계열 일반화 선형 모형의 차수 후보군을 만들고, 차수 후보군의 조합을 이용하여 정보량 기준으로 최종 모형으로 선택한다. 제안된 알고리즘을 평가하기 위하여, 내재적 모형 및 내재적 시계열의 종류에 따른 시뮬레이션 및 실증 분석을 수행하고 예측력을 ARIMA 모형과 비교한다. 예측 성능 평가 결과, 계수형 시계열 분석에서 ARIMA 모형에 비해 시계열 일반화 선형 모형의 예측 성능이 우수함을 확인할 수 있다. 또한 실증분석으로서, 살인사건 발생 건수의 예측결과 ARIMA 모형보다 중기 및 장기 예측에서 우수한 성능을 나타내는 것을 확인할 수 있다.

Keywords

References

  1. Christou, V. and Fokianos, K. (2014). Quasi-likelihood inference for negative binomial time series models, Journal of Time Series Analysis, 35, 55-78. https://doi.org/10.1111/jtsa.12050
  2. Doukhan, P., Fokianos, K., and Tjostheim, D. (2012). On weak dependence conditions for Poisson autoregressions, Statistics & Probability Letters, 82, 942-948. https://doi.org/10.1016/j.spl.2012.01.015
  3. Fahrmeir, L. and Tutz, G. (2001). Multivariate Statistical Modelling based on Generalized Linear Models (2nd ed), Springer, New York.
  4. Ferland, R., Latour, A., and Oraichi, D. (2006). Integer-valued GARCH process, Journal of Time Series Analysis, 27, 923-942. https://doi.org/10.1111/j.1467-9892.2006.00496.x
  5. Fokianos, K., Rahbek, A., and Tjostheim, D. (2009). Poisson autoregression, Journal of the American Statistical Association, 104, 1430-1439. https://doi.org/10.1198/jasa.2009.tm08270
  6. Fokianos, K. and Tjostheim, D. (2011). Log-linear Poisson autoregression, Journal of Multivariate Analysis, 102, 563-578. https://doi.org/10.1016/j.jmva.2010.11.002
  7. Hyndman, R. J. and Khandakar, Y. (2008). Automatic time series forecasting: the forecast package for R, Monash Econometrics and Business Statistics Working Papers 6/07, Monash University, Department of Econometrics and Business Statistics.
  8. Hyndman, R. (2017). Forecast: forecasting functions for time series and linear models, R package version 8.2.
  9. Kedem, B. and Fokianos, K. (2002). Regression Models for Time Series Analysis, John Wiley & Sons, Chichester.
  10. Liboschik, T., Fokianos, K., and Fried, R. (2017). tscount: An R package for analysis of count time series following generalized linear models, Journal of Statistical Software, 82, 1-51.
  11. Tjostheim, D. (2015). Count time series with observation-driven autoregressive parameter dynamics, Handbook of Discrete-Valued Time Series, Handbooks of Modern Statistical Methods, 77-100.
  12. Weiss, C. H. (2008). Thinning operations for modeling time series of counts-a survey, AStA Advances in Statistical Analysis, 92, 319. https://doi.org/10.1007/s10182-008-0072-3