Aptech社国内正規販売代理店　インフォーマティック（株）    [Home]  [Bottom]  [戻る]
                                           動作環境・価格          お問い合わせ・注文
  

     *******************************************
                         Announcing:
                       GAUSS 8.0
          *******************************************

With the introduction of the new sparse matrix data type and the GAUSS
Profiler, the functionality of GAUSS and your ability to optimize your
programs increase dramatically.

Sparse Matrix Data Type

New sparse matrix data type allows for the use of sparse matrices in
many matrix functions and operators, including:

`	|	*	.*
+	-	/	./
/=	./=	==	.==
>	.>	>=	.>=
<	.<	<=	.<=
~	abs	cols	maxc
maxv	minc	minv	print
rows	scalerr	show	type

New Sparse Matrix Functions

The following new functions have been added
for creating and manipulating sparse matrices.

denseToSp	denseToSpRE
packedToSp	spCreate
spDenseSubmat	spDiagRvMat
spEye	spGetNZE
spNumNZE	spOnes
spScale	spSubmat
spToDense	spTrTDense
spTScalar	spZeros

GAUSS Profiler

The new GAUSS Profiler is an important new feature that allows you to optimize your programs rapidly.
The GAUSS Profiler produces a report of how much time your GAUSS pro- grams are spending on
each line and in each called procedure, giving you the information needed to optimize your programs.

Hypotheses Testing Functions

Two new functions in GAUSS implement a new method for testing hypotheses in models with
constraints on parameters described in Constrained Statistical Inference by Mervyn J. Silvapulle and
Pranab K. Sen.
It is well known that current methods for computing standard errors for constrained parameters are only
approximate. These new functions are the only correct method for computing these standard errors, and
hey are only available in GAUSS at this time. ConScore computes the local score statistic for the hypothesis
H(theta) = 0 vs. H(theta) >= 0, where theta is the vector of estimated parameters, and
H() is a constraint function of the parameters.
The model with H(theta) = 0 is estimated, and the Hessian, and optionally the cross-product of
the Jacobian, and the gradient are saved. Then ConScore is called with this information along with the
specification for H(theta) >= 0. The probability of the local score statistic is also computed using a simulation
method employing the quadratic programming solver. ConScore computes both the statistic and its probability.
This statistic has a chi-bar-square distribution. Another GAUSS function, ChiSquareBar, com- putes the
probability of a chi-bar- square distributed statistic given
its covariance matrix and the specification of H(theta) >= 0.

Additional New Features

New in GAUSS Data Archives

Sparse matrix and structure support in GDAﾕ\expnd0s, including support in existing GDA functions
as well as the following new functions:

* gdaGetStructType

* gdaReadSparse

* gdaReadStruct

* Support added for reading from and writing to GDAﾕ\expnd0s created on other platforms, providing you
with an easy and efficient way to transfer data between platforms
* New GDA functions for saving all or a subset of the variables in a workspace to a GDA
(gdaSave) and for loading all of the variables in a GDA into a workspace (gdaLoad)

Other new GDA functions:

* gdaMoment

* gdaOls

New standard deviation functions:

* astd - computes standard deviation of each element along one dimension of anN-dimensional array

* astds- sampleﾕ\expnd0 version of astd, which divides by N rather than N-1

* stdsc- sampleﾕ\expnd0 version of stdc, which divides by N rather than N-1

Support for adding global structures to libraries

Support for extra library paths added to the lib command and the Lib Tool
In GAUSS 7.0, an extra_lib_path variable was added to the GAUSS confi guration file to allow
library statements to fi nd fi les in locations other than the main library path.
Now you may also modify and rebuild libraries located in directories that are included in the extra_lib_path
with the lib command and the Lib Tool

Improved file/line number handling in error returns

* asciiload - loads data stored in an ASCII fi le into GAUSS

* getRow - retrieves a single row from a matrix

* getTrRow - transposes a matrix and then retrieves a single row from it

* maxv - performs an element- by-element comparison of two matrices and returns the maximum value for each element

* minv - performs an element- by-element comparison of two matrices and returns the minimum value for each element

* putvals - inserts multiple values into a matrix

* userErrAt - Prints an error message to the window and error log file, along with the file name and
line number at which the error occurred

* faster intrinsic indsav

Available Platforms:

32-bit: Windows, Linux, Mac
OS X, HP UX11
64-bit: Windows Itanium 2, Linux
AMD, Mac OS X, Solaris

[戻る]

 [Home] [Top]   お問い合わせ・注文