@article{07,
author = "Ette Harrison Etuk and Benjamin Ele Chims",
abstract = "A SARIMA model for forecasting daily
exchange rates of the British Pound (GBP) and the United
States Dollars (USD) is proposed and fitted. The realization
spans from 5th February 2014 to 16th October 2014. The time
plot of the series called herein PDER shows an initial upward
trend up to Saturday 5th July followed by a downward trend.
This means that prior to 5th July the dollar depreciated
relatively after which it appreciated. A seven-day
differencing of PDER yields a series herein called SDPDER
exhibiting a generally horizontal trend. The Augmented
Dickey Fuller unit root test adjudges PDER as non-stationary
and SDPDER as stationary. However the correlogram of
SDPDER shows evidence of non-stationarity in the series. A
non-seasonal differencing of SDPDER yields the series
DSDPDER which also exhibits a horizontal secular trend. It
is certified as stationary by the unit root test. Its correlogram
shows an autocorrelation structure of a stationary series.
Moreover there is an indication of seasonality of period 7
days and the involvement of seasonal autoregressive (AR)
and moving average (MA) components order one each.
Taking into consideration the duality of AR and MA models
suggestive models include (1) A Sarima (0, 1, 1)x(0, 1, 1)7 and
(2) A Sarima (0, 1, 0)x(1, 1, 1)7. On Akaike’s Information
Criterion (AIC) grounds the latter model is the more
adequate. This model further yields a more adequate model:
Sarima(0, 1, 0)x(0, 1, 1)7. Therefore the proposed model is the
Sarima(0, 1, 0)x(0, 1, 1)7. Its residuals are mostly
uncorrelated indicating model adequacy. Forecasting of the
exchange rates may be based on this model",
issn = "2394-2894",
journal = "IJASM",
keywords = "Foreign Exchange Rates;GBP;SARIMA Models;USD",
month = "Sept.- Oct.",
number = "1",
pages = "5-8",
title = "{D}aily {B}ritish {P}ound and {US} {D}ollar {E}xchange {R}ates {M}odelling by {S}easonal {B}ox-{J}enkins {M}ethods",
volume = "1",
year = "2014",
}