Scilab Function
Last update : 16/9/2006
AREG_CAL_LM - Function to calibrate an ARIMA(1,2,1) model on discharge data with a split sample test using lsqrsolve
Calling Sequence
-
[X,CALCcont,CRcont,[CALCcal,CRcal]]=AREG_CAL_LM(Q,NPREV,[ICAL,FOBJ,Xini])
Parameters
-
Q
: Discharge data (matrix [Nx1] where N is the number of time steps)
-
NPREV
: List of forecast lead-time (expressed in number of time-steps, matrix [NPx1] with NP the number of forecast lead-time)
-
ICAL
: Indices of the value to use in the Q vector during the calibration process
-
FOBJ
: Transformation applied to calculate the objective function (default = 'Q', possible = 'logQ' - logarithm transform - or 'racQ' -square root transform -)
-
X
: Optimised parameters (matrix [3x3]). X(:,1) are the optimised parameters on P1, X(:,2) on P2 and X(:,3) on P3.
-
CALCcont
: Discharge forecasts in validation mode (matrix [N x (NP+1)]). CALCcont(:,1) is equal to Q, CALCcont(:,2) are the forecasts for the first
lead-time defined in NPREV vector ,...
-
CRcont
: Quality criteria in validation mode (matrix [2 x NP]). For each lead-time, 1 persistance (CRcont(1,1)) and 1 RMSE (CRcont(2,1:2)) are calculated.
The first with perfect rainfall and the second with null rainfall.
Description
-
The function performs the following actions:
-
(1) The dataset is split into 2 equal parts, P1 and P2. P3 represents the whole dataset.
-
(2) On each sub-period, an ARIMA(1,2,1) is calibrated with the lsqrsolve.
-
(3) The model quality in validation mode is assessed by calculating persistance criteria on a simulation made on P1 with parameters from P2 and vice-versa.
See Also
c_AREG
,
PasaPas
,
Authors
-
Julien Lerat
CEMAGREF Antony, HBAN Unit, julien.lerat@cemagref.fr