gets-package: General-to-Specific (GETS) Modelling and Indicator Saturation...

Description Details Author(s) References See Also Examples

Description

Automated multi-path General-to-Specific (GETS) modelling of the mean and variance of a regression, and indicator saturation methods for detecting structural breaks in the mean. The mean can be specified as an autoregressive model with covariates (an 'AR-X' model), and the variance can be specified as a log-variance model with covariates (a 'log-ARCH-X' model).

The four main functions of the package are arx, getsm, getsv and isat. The first function, arx, estimates an AR-X model with (optionally) log-ARCH-X errors. The second function, getsm, undertakes GETS model selection of the mean specification of an arx object. The third function, getsv, undertakes GETS model selection of the log-variance specification of an arx object. The fourth function, isat, undertakes GETS model selection of an indicator saturated mean specification.

The package also provides auxiliary functions used by the main functions, in addition to extraction functions (mainly S3 methods).

Details

Package: gets
Type: Package
Version: 0.13
Date: 2017-10-05
License: GPL-2

The code originated in relation with G. Sucarrat and A. Escribano (2012): 'Automated Financial Model Selection: General-to-Specific Modelling of the Mean and Volatility Specifications', Oxford Bulletin of Economics and Statistics 74, Issue 5 (October), pp. 716-735. Subsequently, Felix Pretis and James Reade joined for the development of the isat code and related functions

Author(s)

Felix Pretis, http://www.felixpretis.org/
James Reade, https://sites.google.com/site/jjamesreade/
Genaro Sucarrat, http://www.sucarrat.net/

Maintainer: Genaro Sucarrat

References

G. Sucarrat and A. Escribano (2012): 'Automated Financial Model Selection: General-to-Specific Modelling of the Mean and Volatility Specifications', Oxford Bulletin of Economics and Statistics 74, Issue 5 (October), pp. 716-735

Carlos Santos, Hendry, David, F. and Johansen, Soren (2007): 'Automatic selection of indicators in a fully saturated regression'. Computational Statistics, vol 23:1, pp.317-335

Jurgen, A. Doornik, Hendry, David F., and Pretis, Felix (2013): 'Step Indicator Saturation', Oxford Economics Discussion Paper, 658.

See Also

arx, getsm, getsv, isat

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
##Simulate from an AR(1):
set.seed(123)
y <- arima.sim(list(ar=0.4), 60)

##Estimate an AR(2) with intercept as mean specification
##and a log-ARCH(4) as log-volatility specification:
myModel <- arx(y, mc=TRUE, ar=1:2, arch=1:4)

##GETS modelling of the mean of myModel:
simpleMean <- getsm(myModel)

##GETS modelling of the log-variance of myModel:
simpleVar <- getsv(myModel)

##results:
print(simpleMean)
print(simpleVar)

##step indicator saturation of an iid normal series:
set.seed(123)
y <- rnorm(30)
isat(y)

Example output

Loading required package: zoo

Attaching package: 'zoo'

The following objects are masked from 'package:base':

    as.Date, as.Date.numeric

2 paths to search
Searching: 1 2 
3 paths to search
Searching: 1 2 3 

Date: Wed Aug  9 13:20:30 2017 
Dependent var.: y 
Method: Ordinary Least Squares (OLS)
Variance-Covariance: Ordinary 
No. of observations (mean eq.): 58 
Sample: 3 to 60 

GUM mean equation:

       reg.no keep     coef std.error  t-stat  p-value
mconst      1    0 0.026155  0.117029 0.22349 0.823979
ar1         2    0 0.333996  0.134355 2.48593 0.015988
ar2         3    0 0.034490  0.133176 0.25898 0.796615

GUM log-variance equation:

            coef std.error  t-stat p-value
vconst -0.278502  0.391091  0.5071 0.47639
arch1   0.215508  0.139456  1.5453 0.12870
arch2   0.105365  0.142566  0.7391 0.46340
arch3   0.081331  0.144165  0.5642 0.57522
arch4  -0.306401  0.140825 -2.1758 0.03443

Diagnostics:

                  Chi-sq df p-value
Ljung-Box AR(3)   2.4462  3 0.48509
Ljung-Box ARCH(5) 1.3238  5 0.93246
Jarque-Bera       2.0576  2 0.35744

Paths searched: 

path 1 : 1 3 
path 2 : 3 1 

Terminal models: 

spec 1 : 1 2 3 
spec 2 : 2 

              info(sc)     logl  n  k
spec 1 (gum):   2.6473 -65.4945 54  3
spec 2:         2.5464 -66.7582 54  1

SPECIFIC mean equation:

       coef std.error t-stat  p-value
ar1 0.34686   0.12327 2.8138 0.006708

SPECIFIC log-variance equation:

            coef std.error  t-stat p-value
vconst -0.212822  0.473616  0.2019  0.6532
arch1   0.194178  0.142006  1.3674  0.1777
arch2   0.070728  0.144627  0.4890  0.6270
arch3   0.041949  0.145447  0.2884  0.7742
arch4  -0.207758  0.142741 -1.4555  0.1519

Diagnostics:

                   Chi-sq df p-value
Ljung-Box AR(3)   2.66621  3 0.44600
Ljung-Box ARCH(5) 0.50528  5 0.99193
Jarque-Bera       1.61710  2 0.44550
                          
SE of regression   0.87620
R-squared          0.12105
Log-lik.(n=54)   -66.75816

Date: Wed Aug  9 13:20:30 2017 
Method: Ordinary Least Squares (OLS)
No. of observations (variance eq.): 54 
Sample: 7 to 60 

GUM log-variance equation:

       reg.no keep      coef std.error   t-stat  p-value
vconst      1    1 -0.278502  0.391091  0.50711 0.476392
arch1       2    0  0.215508  0.139456  1.54534 0.128698
arch2       3    0  0.105365  0.142566  0.73906 0.463396
arch3       4    0  0.081331  0.144165  0.56415 0.575224
arch4       5    0 -0.306401  0.140825 -2.17576 0.034425

Diagnostics:

                  Chi-sq df p-value
Ljung-Box AR(3)   2.4462  3 0.48509
Ljung-Box ARCH(5) 1.3238  5 0.93246
Jarque-Bera       2.0576  2 0.35744

Paths searched: 

path 1 : 2 4 3 5 
path 2 : 3 4 2 5 
path 3 : 4 3 2 5 

Terminal models: 

spec 1 : 1 2 3 4 5 
spec 2 : 1 

                info(sc)     logl  n  k
spec 1 (gum):     2.7951 -65.4945 54  5
spec 2 (empty):   2.6591 -69.8006 54  1

SPECIFIC log-variance equation:

           coef std.error t-stat p-value
vconst -0.25267   0.26409 0.9154  0.3387

Diagnostics:

                   Chi-sq df p-value
Ljung-Box AR(3)   1.50038  3 0.68218
Ljung-Box ARCH(5) 3.88666  5 0.56585
Jarque-Bera       0.38472  2 0.82501
                          
SE of regression   0.88132
R-squared         -0.00194
Log-lik.(n=54)   -69.80062

SIS block 1 of 2:

15 paths to search
Searching: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 

SIS block 2 of 2:

14 paths to search
Searching: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 

GETS of union of retained SIS variables... 

GETS of union of ALL retained variables...


Date: Wed Aug  9 13:20:30 2017 
Dependent var.: y 
Method: Ordinary Least Squares (OLS)
Variance-Covariance: Ordinary 
No. of observations (mean eq.): 30 
Sample: 1 to 30 

GUM mean equation:

       reg.no keep      coef std.error   t-stat p-value
mconst      1    1 -0.047104  0.179111 -0.26299 0.79442

Diagnostics:

                    Chi-sq df p-value
Ljung-Box AR(1)   0.046575  1 0.82913
Ljung-Box ARCH(1) 3.367118  1 0.06651
Jarque-Bera       0.609541  2 0.73729

Paths searched: 

NULL

Terminal models: 

spec 1 : 1 

              info(sc)     logl  n  k
spec 1 (gum):   2.8796 -41.4936 30  1

SPECIFIC mean equation:

              coef std.error     t-stat   p-value
mconst -0.04710376 0.1791109 -0.2629866 0.7944204

Diagnostics:

                    Chi-sq df p-value
Ljung-Box AR(1)   0.046575  1 0.82913
Ljung-Box ARCH(1) 3.367118  1 0.06651
Jarque-Bera       0.609541  2 0.73729
                          
SE of regression   0.98103
R-squared          0.00000
Log-lik.(n=30)   -41.49361

gets documentation built on Oct. 8, 2017, 1:03 a.m.