Issue 64, 19 July 2007

In this issue:

Top Story - Partial Least Squares

Regression by means of projections to latent structures (PLS, also known as partial least squares) is a useful alternative to the linear multiple regression model fitted by 'least squares' if:

  • the number of x-variables is relatively high compared with the number of observations;
  • the x-variables are correlated;
  • there is more than one y-variable (response variable) and these variables are correlated.

Hence the PLS method is popular in industries that collect correlated data on many x-variables, known as predictors. For example, multivariate calibration in analytical chemistry; spectroscopy in chemometrics; and quantitative structure activity relationships (QSAR) in drug design.

The PLS method extracts orthogonal linear combinations of predictors, known as factors, from the predictor data that explain variance in both the predictor variables and the response variable(s). For the full article, go to

Product News - Using the NAG Toolbox for MATLAB® - Part 2

As reported in Issue 61 ( of NAGNews, the new NAG Toolbox for MATLAB® is now available for beta test by NAG Library users. The NAG Toolbox makes the full functionality of the NAG Library available through MATLAB, thereby enhancing the basic functionality that MATLAB has to offer in many areas. An advantage of calling NAG via MATLAB is that many routine arguments become optional or unnecessary, which makes code easier to read and maintain. Moreover, all of the NAG Library documentation has been converted to MATLAB help format, which makes it simple to access via MATLAB's usual documentation facilities.

An earlier article ( discussed some of the contents of the NAG Toolbox, looking at special functions, surface fitting and function minimization. In this second article, we continue our exploration of the NAG Toolbox, paying particular attention to root finding, curve fitting, interpolation and the creation and use of a copula in modelling. For the second article, go to

General Information - A Practical Test for the NAG Optimisation Routines

Jan Stebel, Charles University, Prague.

In many real-life applications of the NAG library a physical situation is first modelled, often with some unknown parameters, and the library used to solve the model and determine the best values for the unknown parameters.

In the article Jan Stebel describes one such application, the determination of an optimal shape for the back wall of a paper-making machine header.

Forthcoming Events - US, UK
  • Webinar - Rapid Financial Model Development using NAG Routines and the Maple 11 Environment
    Wednesday 22 August 10am EDT / 3pm BST
    This webinar will demonstrate how financial models can be efficiently and effectively developed by calling NAG routines from within the Maple 11 environment. Webinar examples will range from posing, visualizing, and answering simple financial-based questions to developing sophisticated pricing models using both analytic and numeric approaches. All examples will demonstrate how model development is facilitated by using the proper environment (one that is well-suited for knowledge capture and deployment) in addition to highly optimized code. For further information on this event, contact us at .

  • IMA Conference on Numerical Linear Algebra and Optimisation
    13-15 September 2007, University of Birmingham
    The success of modern codes for large-scale optimisation is heavily dependent on the use of effective tools of numerical linear algebra. On the other hand, many problems in numerical linear algebra lead to linear, nonlinear or semidefinite optimisation problems. The purpose of the conference is to bring together researchers from both communities and to find and communicate points and topics of common interest. This is the first IMA Conference of this theme, and NAG is pleased to be present. For further information visit

  • The Mathematical Association Secondary Education Mathematics Conference
    15 September 2007, Stirling University
    The Association exists to bring about improvements in the teaching of mathematics and its applications, and to provide a means of communication among students and teachers of mathematics. This conference will provide a forum for those involved in the teaching of mathematics to meet, and NAG is delighted to be supporting it. For further information on this conference, visit .

Tips & Hints - Kusari Licence Management System

Some of our users may have noticed that we are gradually moving to a new licence management system, called Kusari.

As we move to this system we have with 3 different systems in the field:
a) Kusari
b) FLEXlm
c) A PC-specific system

Eventually all our products will use Kusari and licence key installation will become uniform once more. We are, in addition, developing an interface to make the installation of a licence key much simpler.

Because many of our users are single-user, PC-Windows based, this note gives an overview of the steps required to obtain and install a Kusari licence key on this system. Other machine ranges will follow the same general procedures, but will differ in detail. The more detailed information is, in all cases, found under the license subdirectory or from or in the Installer's Note for the appropriate implementation. These are provided with the product but are also available on the web (see

If you completed the ‘trial request form’ on the NAG download page then you should have received a trial key. If so you may jump straight to stage d); otherwise begin at a)

a) Obtain a machine identification string using the khostid program provided. A good technical tip here is to save the output from this program to a file. (On a PC this means opening a Command Prompt, changing to the .\license\bin\w32 subdirectory and typing khostid>nag.txt at the command prompt.)

b) Edit the file nag.txt, perhaps using Notepad or otherwise your favourite plain text editor, to include the product code of the product you are installing, for example FLDLL214Z if you are installing the Mark 21 Fortran DLLs. Also include your name, the company name and address and the NAG site reference, if known.

c) Send the file by email to your local NAG office. This will be one of, or

d) You will receive a licence key back from your local office. Place this in a file, or append it to an existing Kusari licence file (but not a FLEXlm licence file), making sure that there is a blank line at the end of the file. As with b), if you are editing an existing file, or creating a new file, then a plain text editor, such as Notepad, must be used. Products such as Word, which are not plain text editors, are NOT suitable for this purpose.

e) The environment variable NAG_KUSARI_FILE now needs to be set. This is done via START/Control Panel/System/Advanced/Environment Variables. From here create NAG_KUSARI_FILE if this environment variable does not exist. Append to the value of the environment variable the full hierarchical name of the file containing the licence. For example NAG_KUSARI_FILE might be set to just c:\program files\nag\licenses\naglic.txt if NAG_KUSARI_FILE did not exist before. Otherwise for example it might end up set to C:\program_files\naglic1.txt;c:\program files\nag\licenses\naglic.txt

In the latter case an alternative approach might have been to append the new licence key to the end of the licence file that already exists. (In the example above this would be C:\program_files\naglic1.txt.)

It is important to note that the licence management software searches only until a matching licence is recognized. If, for example, an expired licence is found in C:\program_files\naglic1.txt, the search will stop at this point even though a fresh licence is available in c:\program files\nag\licenses\naglic.txt. We therefore recommend that expired licences be removed from licence files as the licences expire.

f) Finally check that the licence is correctly installed by running a00acf example program. The Mark 21 DLLs have a program NAG_Fortran_DLL_info.exe in the Diagnostic subdirectory. This may also be run.

All previous Tips & Hints can be found in the NAG Tips & Hints Repository at

Product News - New NAG Library Implementations

NAG is committed to offering new implementations of its broad range of numerical and statistical software components and compilers and tools. The following implementations for NAG products have become available since the last issue of our newsletter:

The NAG Fortran Library (, Mark 21 is now also available for the following platforms:

  • x86-32, Linux, gfortran compiler, Double Precision
  • AMD64/EM64T, Linux 64, gfortran compiler, Double Precision

Please note that the above products will work using versions 4.2 or 4.3 (experimental) of the compiler. They will NOT work with earlier versions of the compiler (such as version 4.1.2).

The NAG Parallel Library (, Release 3 is now also available for the following platform:

  • Sun UltraSPARC 64 bit, Solaris, Sun Fortran, Sun ClusterTools 7

The NAGWare Compiler (, Release 5.1 has been re-issued for the following platforms:

  • Linux (x86) glibc2.3 or later
  • AMD64/EM64T Linux glibc2.3 or later
For full details of these and all other available implementations, visit the NAG site. Comprehensive technical details of each implementation are given in the relevant Installation and User Notes at

Look out for announcements of the following implementations in future issues of NAGNews:

  • The NAGWare Compiler for Apple Intel Mac (expected soon)
  • The NAG Fortran Library for 64-bit Windows (expected soon)
  • The NAG SMP Library for Sun/Opteron
  • The NAG C Library for 64-bit Windows
  • The NAGWare compiler for 32-bit Windows
Intel 10.0 compilers for Windows and the NAG products

Intel® has now released version 10.0 of both C and Fortran compilers for Windows. The latest NAG Windows products have been tested using these compilers.

The two NAG Fortran Libraries (
  • FLW3221DEL - Windows NT/2000/XP, Intel Visual Fortran (/MT)
  • FLDLL214ZL - Windows NT/2000/XP Dynamic Link Library
and the NAG Fortran 90 Library (
  • FNW3204DCL - Windows NT/2000/XP, Intel Visual Fortran
have been tested using the Intel compiler
Intel® Visual Fortran Compiler for applications running on IA-32
Version 10.0 - Build 20070426 Package ID: W_FC_P_10.0.025
The results have been compared with the distributed results. The few differences are acceptable.

The NAG C Library product (
  • CLW3208DAL - Windows XP/2003, Microsoft C/C++ (Static and plain DLL)
Although this product was built using Microsoft C/C++ compiler, it has been tested using the Intel compiler:
Intel® C++ Compiler for applications running on IA-32, Version 10.0
Build 20070426 Package ID: W_CC_P_10.0.025
The results have been compared with the distributed results. The few differences are acceptable.

In Summary
  • FLW3221DEL, FLDLL214ZL and FNW3204DCL will work using version 10.0 of the Intel Fortran compiler for Windows (Package ID: W_FC_P_10.0.025) and compatible.
  • CLW3208DAL will work using version 10.0 of the Intel C++ compiler for Windows (Package ID: W_CC_P_10.0.025) and compatible.

Managing your subscription

Please feel free to forward this newsletter to colleagues or to post it to your company's Intranet sites.

To subscribe, unsubscribe or learn more about managing your subscription, please see: