Abstract
The adaptive lasso can consistently identify the true model in regression model. However, the adaptive lasso cannot account for lag effects, which are essential for a time series model. Consequently, the adaptive lasso can not reflect certain properties of a time series model. To improve the forecast accuracy of a time series model, we propose a lag weighted lasso. The lag weighted lasso imposes different penalties on each coefficient based on weights that reflect not only the coefficients size but also the lag effects. Simulation studies and a real example show that the proposed method is superior to both the lasso and the adaptive lasso in forecast accuracy.
Similar content being viewed by others
References
Efron B, Hastie T, Johnstone I, Tibshirani R (2004) Least angle regression. Ann Stat 32: 407–499
Efron B, Hastie T, Tibshirani R, Friedman J (2008) The elements of statistical learning. Springer, Berlin
Fan J, Li R (2001) Variable selection via nonconcave penalized likelihood and its oracle properties. J Am Stat Assoc 96: 1348–1360
Lee JH (2007) Time series analysis and application. Freedom Academy, Gyeonggi-do, Korea
Matsumoto A, Szidarovszly F (2010) Dynamic goodwin’s business cycles with fized and continuously distributed time delays
Park HW (2009) Effects of environmental factors on monthly cerebrovascular mortality in Seoul, Korea. J Korean Data Anal Soc 11: 687–698
Pesaran MH (1997) An autoregressive distributed lag modelling approach to cointegration analysis. Science 7825(94): 371–413
Ravines RR, Schmidt RM, Migon HS (2006) Revisiting distributed lag models through a Bayesian perspective. J Appl Stoch Midels Bus Ind 22: 193–210
Shrestha SL (2007) Time series modeling of respiratory hospital admissions and geometrically weighted distributed lag effects from ambient particulate air pollution within Kathmandu Valley, Nepal. Environ Model Assess 12: 239–251
Tibshirani R (1996) Regression shrinkage and selection via the lasso. J R Stat Soc 58: 267–288
Tibshirani R (2006) Revisiting distributed lag models through a bayesian perspective. J Appl Stoch Models Bus Ind 22: 193–210
Wang H, Leng C (2008) A note on adaptive group lasso. Comput Stat Data Anal 52: 5277–5286
Yuan M, Lin Y (2007) On the nonnegative garrote estimator. J R Stat Soc 69: 143–161
Zou H (2006) The adapitve lasso and its oracle properties. J Am Stat Assoc 101: 1418–1429
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Park, H., Sakaori, F. Lag weighted lasso for time series model. Comput Stat 28, 493–504 (2013). https://doi.org/10.1007/s00180-012-0313-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00180-012-0313-5