<< fernandez1 Time series disaggregation litterman1 >>

Grocer >> Time series disaggregation > litterman

litterman

Temporal disaggregation using the Litterman method

CALLING SEQUENCE

[y,res]=litterman(namey,arg1,...,argn)

PARAMETERS

Input

* namey = a time series, a real (n x 1) vector or a string equal to the name of a time series or a (n x 1) real vector between quotes, representing the low frequency data that must be desaggregated

* argi = an argument which can be

  - a time series

  - a real (n x k) matrix

  - a string matrix whose elements represent the names of a time series or a (nx1) real vector between quotes

  - the string 'divfq=n' where n is the number of high frequency data points for each low frequency data points

  - the string 'typemin=xxx' where xxx is the maximisation method (llike -default- or wls)

  - the string 'ta=n' where n is the aggregation type:

    n=-1 (default) ---> sum (flow)

    n=0 ---> average (index)

    n=i ---> i th element (stock) ---> interpolation

  -the string 'delta=x' where x is the increment used for numerical derivation 

- 'optfunc=optim' if the user wants to use the optim optimisation function (default: optimg)

- 'opt_nelmead=crit,nitermax' with crit the value of the convergence criterion in the Nelder-Meade optimisation function and nitermax the maximum number of iterations (default = 'opt_nelmead=2*%eps,1000')

- 'opt_optim=opts' where opts are options for optim that can be entered after the starting value of the parameters (default = 'opt_optim=,''ar'',1e6,1e6'')

- 'opt_convg=val' where val is the threshold on gradient norm (default = 'opt_convg=1e-5')

 

Output

* y = High frequency estimate

* res = a results tlist with:

   - res('meth') = 'Litterman'

   - res('typemod') = type of the model for the high frequency innovations

   - res('ta') = type of disaggregation

   - res('nobs_lf') = nobs. of low frequency data

   - res('nobs_hf') = nobs. of high-frequency data

   - res('pred') = number of extrapolations

   - res('s') = frequency conversion between low and high freq.

   - res('p') = number of regressors (including intercept)

   - res('y') = high frequency estimate

   - res('y_lf') = low frequency data

   - res('indicator') = high frequency indicators

   - res('y_dt') = high frequency estimate: standard deviation

   - res('y_up') = high frequency estimate: sd + sigma

   - res('y_lo') = high frequency estimate: sd - sigma

   - res('resid') = high frequency residuals

   - res('resid_lf') = low frequency residuals

   - res('beta') = estimated model parameters

   - res('sd') = standard deviation of the estimated model parameters

   - res('tstat') = estimated model parameters: t ratios

   - res('rho') = innovational parameter

   - res('aic') = Information criterion: AIC

   - res('bic') = Information criterion: BIC

   - res('llike') = Objective function used by the estimation method

   - res('typemin') = method of estimation

   - res('llike') = Log-likelihood at the estimated parameters

   - res('sigma') = Variance at the estimated parameters

   - res('namey') = Name of the low frequency data

   - res('namex') = Name of the indicators

DESCRIPTION

Temporal disaggregation using the Litterman method of temporal disaggregation (high level function with vectors, matrices or ts and the possibility of default parameters).

EXAMPLE

load(GROCERDIR+'\data\xesp.dat')
[y,res] = litterman('Y','x','ta=-1','typemin=wls');
// Example taken from function litterman_d. Provides quarterly disaggregation
// of (22 x 1) matrix Y (representing annual Spanih exportations) with (88 x 1) matrix x.
// Annual series are built by summing quarterly series (ta=-1); the order of aggregation is determined by the function;
// the minimisation applies to weighted least squares ('typemin=wls').

AUTHOR

Eric Dubois 2005

Report an issue
<< fernandez1 Time series disaggregation litterman1 >>