Добавил:
Upload Опубликованный материал нарушает ваши авторские права? Сообщите нам.
Вуз: Предмет: Файл:
Eviews5 / EViews5 / Docs / EViews 5 Command Ref.pdf
Скачиваний:
91
Добавлен:
23.03.2015
Размер:
5.23 Mб
Скачать

494—Appendix B. Command Reference

deriv=keyword Set derivative methods. The argument keyword should be a oneor two-letter string. The first letter should either be “f” or “a” corresponding to fast or accurate numeric derivatives (if used). The second letter should be either “n” (always use numeric) or “a” (use analytic if possible). If omitted, EViews will use the global defaults.

p

Print estimation results.

Examples

sys1.sur(i)

estimates SYS1 by SUR, iterating simultaneously on the weighting matrix and coefficient vector.

nlsys.sur(showopts,m=500)

estimates NLSYS by SUR with up to 500 iterations. The “showopts” option displays the starting values.

Cross-references

See Chapter 23, “System Estimation”, on page 693 of the User’s Guide for a discussion of system estimation.

svar

Var Proc

 

 

Estimate factorization matrix for structural innovations.

Syntax

Var Proc:

var_name.svar(options)

The var object must previously have been estimated in unrestricted form.

You must specify the identifying restrictions either in text form by the append proc or by a pattern matrix option. See “Specifying the Identifying Restrictions” on page 731 of the User’s Guide for details on specifying restrictions.

Options

You must specify one of the following restriction type:

svar—495

rtype=text

Text form restrictions. The restrictions must be specified

 

by the append command to use this option.

 

 

rtype=patsr

Short-run pattern restrictions. You must provide the

 

names of the patterned matrices by the “namea=” and

 

“nameb=” options as described below.

 

 

rtype=patlr

Long-run pattern restrictions. You must provide the

 

name of the patterned matrix by the “namelr=” option

 

as described below.

 

 

Other Options:

namea=arg,

Names of the pattern matrices for A and B matrices.

nameb=arg

Must be used with “rtype=patsr”.

 

 

namelr=arg

Name of the pattern matrix for long-run impulse

 

responses. Must be used with “rtype=patlr”.

 

 

fsign

Do not apply the sign normalization rule. Default is to

 

apply the sign normalization rule whenever applicable.

 

See “Sign Indeterminacy” on page 736 of the User’s

 

Guide for a discussion of the sign normalization rule.

 

 

f0=arg

Starting values for the free parameters: “scalar” (specify

(default=0.1)

fixed value for starting values), “s” (user specified start-

 

ing values are taken from the C coefficient vector), “u”

 

(draw starting values for free parameters from a uni-

 

form distribution on [0,1]), “n” (draw starting values

 

for free parameters from standard normal).

 

 

maxiter=inte-

Maximum number of iterations. Default is taken from

ger

global option setting.

 

 

conv=number

Convergence criterion. Default is taken from global

 

option setting.

 

 

trace=integer

Trace iterations process every integer iterations (dis-

 

plays an untitled text object containing summary infor-

 

mation).

 

 

nostop

Suppress “Near Singular Matrix” error message even if

 

Hessian is singular at final parameter estimates.

Examples

var var1.ls 1 4 m1 gdp cpi

matrix(3,3) pata

496—Appendix B. Command Reference

'fill matrix in row major order

pata.fill(by=r) 1,0,0, na,1,0, na,na,1

matrix(3,3) patb

pata.fill(by=r) na,0,0, 0,na,0, 0,0,na

var1.svar(rtype=patsr,namea=pata,nameb=patb)

The first line declares and estimates a VAR with three variables. Then we create the shortrun pattern matrices and estimate the factorization matrix.

var var1.ls 1 8 dy u @

var1.append(svar) @lr1(@u1)=0

freeze(out1) var1.svar(rtype=text)

The first line declares and estimates a VAR with two variables without a constant. The next two lines specify a long-run restriction in text form and stores the estimation output in a table object named OUT1.

Cross-references

See “Structural (Identified) VARs” on page 731 of the User’s Guide for a discussion of structural VARs.

sym

Object Declaration

 

 

Declare a symmetric matrix object.

The sym command declares and optionally initializes a matrix object.

Syntax

Command: sym(n) sym_name[=assignment]

sym takes an optional argument n specifying the row and column dimension of the matrix and is followed by the name you wish to give the matrix.

You may also include an assignment in the sym command. The sym will be resized, if necessary. Once declared, symmetric matrices may be resized by repeating the sym command for a given matrix name.

Examples

sym mom

declares a symmetric matrix named MOM with one zero element.

table—497

sym y=@inner(x)

declares a symmetric matrix Y and assigns to it the inner product of the matrix X.

Cross-references

See “Matrix Language” on page 23 for a discussion of matrix objects in EViews.

See also matrix (p. 366).

system

Object Declaration

 

 

Declare system of equations.

Syntax

Command: system system_name

Follow the system keyword by a name for the system. If you do not provide a name, EViews will open an untitled system object (if in interactive mode).

Examples

system mysys

creates a system named MYSYS.

Cross-references

Chapter 23, “System Estimation”, on page 693 of the User’s Guide provides a full discussion of system objects.

See append (p. 205) for adding specification lines to an existing system. The system object is described in greater detail in “System” (p. 184).

table

Object Declaration

 

 

Declare a table object.

The table command declares and optionally sizes a table object.

Syntax

Command: table(rows,cols) table_name

Соседние файлы в папке Docs