Iheanyi S. Iwueze
Department of Statistics, Abia State University, P.M.B. 2000, Uturu, Abia State, Nigeria
Anthony C. Akpanta
Department of Statistics, Abia State University, P.M.B. 2000, Uturu, Abia State, Nigeria
ABSTRACT
In this study we present some solutions to the problems encountered when a time series multiplicative model is linearized by taking the natural logarithm of the observed series. Problems associated with the trend-cycle component, with a view to achieving no alteration (if possible) to the fundamental nature of the original data, are discussed for a class of trending curves which includes the linear, quadratic and exponential growth curves. Necessary and sufficient conditions are derived for the parameters of the linear and quadratic curves. Numerical examples are given to illustrate the results obtained.
PDF References Citation
How to cite this article
Iheanyi S. Iwueze and Anthony C. Akpanta, 2007. Effect of the Logarithmic Transformation on the Trend-cycle Component. Journal of Applied Sciences, 7: 2414-2422.
DOI: 10.3923/jas.2007.2414.2422
URL: https://scialert.net/abstract/?doi=jas.2007.2414.2422
DOI: 10.3923/jas.2007.2414.2422
URL: https://scialert.net/abstract/?doi=jas.2007.2414.2422
REFERENCES
- Chaddha, A.L. and S.S. Chitgopekar, 1971. A generalization of the logistic curves and long-range forecasts. Bell J. Busin. Manage. Sci., 2: 542-560.
Direct Link - Iwueze, I.S. and E.C. Nwogu, 2004. Buys-Ballot estimates for time series decomposition. Global J. Math. Sci., 3: 83-98.
Direct Link - Iwueze, I.S., 2007. Some implications of truncating the N(1, σ2) distribution to the left at zero. J. Applied Sci., 7: 189-195.
CrossRefDirect Link - Osborne, J., 2002. Notes on the use of data transformations. J. Practic. Assess. Res. Evalu., 8: 1-7.
Direct Link