# NAGnews 118 | 28 November 2013

Posted on
29 Nov 2013

In this issue

New services to underpin and innovate your numerical computing revealed by NAG

NAG attended the 25th Supercomputing conference in Denver last week and took this occasion - NAG has had an exhibit at every single SC event - to announce its new Numerical Services. For over four decades NAG has been at the forefront of the development of tried and trusted numerical algorithms, in the form of the NAG Libraries. In response to customer interest NAG will now deliver the expertise and experience behind the highly reputable Libraries and Compiler directly to developers and users of numerical computing applications. This follows the success of NAG's HPC Services and enables customer access to the full range of NAG experts - numerical analysts, computer scientists, mathematicians, algorithm developers, software engineers and more.

Recent successes of the NAG Numerical Services include:

• Helping a pharmaceutical company deliver a tool to run scenario analysis on complete product lifecycles with robust and tested methods
• A consumer products firm, with a focus on pricing models, used NAG to access some sophisticated regression techniques
• The financial systems division of a technology services company needed NAG to enhance key statistical models
• An analysis and pricing firm used NAG to help speed up their application by x 10

New technical tip: Using the NAG Compiler with the NAG Fortran Library (Mark 24) on Windows

The NAG Fortran Compiler is an excellent compiler for checking and running your Fortran code. We use it extensively here at NAG to ensure that our code for the library complies with the current Fortran standards.

Personally whenever I have a user problem report that I can't resolve by inspection my first instinct is to run the user's code with the Compiler. Frequently this identifies the error immediately.

As I am a Windows user, I am able to make use of the Integrated Development Environment (IDE) for the compiler that is provided to our Windows users. We call this IDE 'NAG Fortran Builder'. One of the nice features of this IDE is its ease of use with the prevailing NAG libraries. To do this the user normally specifies a 'NAG Library Project' at creation time; thereafter the relevant settings are made to the compiler, so that either the Windows 64-bit or Windows 32-bit DLLs are used.

At points in the NAG cycle a situation arises where the latest Fortran Builder is released. It automatically picks up the then current NAG libraries. Information about these libraries is embedded within Fortran Builder. This information includes documentation, interface blocks and example program information, all of which is mark-dependent. Subsequently a new NAG Library is released and so our Windows users would want to use the latest library from Fortran Builder. Some guidance is given on this within the relevant Users' Note. This note attempts to gather together and expand on this information.

Read the whole story on the NAG Blog here.

Do you need to perform the Dickey-Fuller test?

We are often asked about new functionality and many customer requests end up in new releases of the NAG Library. The last edition of NAGnews featured one such request: Change Point Analysis.

This month we highlight another: the Dickey-Fuller test. If you would like to discuss taking an early code release or are just interested in sharing information on how you use such an algorithm simply reply to this email.

In a similar vein, if you have a particular requirement for a routine or multiple routines that don't feature in the Library do let us know.

TakeAIM: Articulating the influence of mathematics

NAG is proud to sponsor the TakeAIM competition organized by the Industrial Mathematics Knowledge Transfer Network. The goal of TakeAIM is to highlight the crucial role that mathematics plays and will increasingly play in all aspects of our lives through the voice of mathematics students. David Sayers, Principal Technical Consultant at NAG was part of the judging panel for the competition and was really impressed at the quality of entries. The winners were awarded their prize at an awards ceremony held in Oxford on 25th November.

Congratulations to the winners:

Francis Watson, University of Manchester, Better imaging for landmine detection
and
Leonor Garcia Gutierrez, University of Warwick, Maths meets hypertension.

The two runners up were:

Anna Cederlund, University of Aberdeen, The Cartilage Calculator
and
Ellen Webborn, University of Warwick, The mathematics of electricity storage.

Mark 24 new functionality spotlight: Non-negative least squares problems

Optimization solutions for Non-Negative Least Squares problems (Bounded-Variable Least Squares)

A search on the Web will quickly reveal numerous applications for a routine which finds the best fit vector x to a system of linear equations where the components of x are constrained to be non-negative. For example statisticians may wish to fit a linear regression model:

y = Xc + e


where y is a vector of observations, X a design matrix and e a vector of noise. If c represents inherently non-negative quantities, such as prices, growth rates or pixel intensity perhaps, then a constraint might be that the elements of c be non-negative. Such a problem is termed a non-negative (linear) least squares problem.

We may similarly find situations where, additionally, the 'solution' should not have components exceeding a given value. Such a requirement defines a bounded-variable (linear) least squares problem.

The NAG optimization chapters contain very powerful and general routines, so such problems could always be cast into a form where the problems could be addressed by these routines. When we did this at the request of a user we found, not surprisingly, that these non-specialist routines were not as efficient as a routine designed specifically for such problems.

Consequently, we decided to add to the functionality of the NAG Library, at Mark 24, by including a specialist routine for bounded-variable least squares problems. This routine is E04PCx and is designed for problem sizes that are not too big. (Dense matrices and techniques are employed throughout. Depending upon user-demand, we will consider specialist sparse algorithms for larger problems at future marks.)

Read the full Mark 24 mini-article here.

Call for papers: Bachelier Finance Society World Congress

The 8th World Congress of the Bachelier Finance Society will take place in Brussels, Belgium from 2-6 June 2014. Each year the congress attracts hundreds of participants from quantitative and mathematical finance.

NAG are delighted to be involved in this event and are assisting the Society with their Call for Papers.

Abstract submission deadline (by 31 December 2013):

Abstracts must be submitted online via this link. An author may be presenter of only one presentation. The following required information on the abstract needs to be completed:

• Abstract title (max. 160 characters)
• Personal data of the corresponding author
• Presentation format preference (oral presentation or poster presentation)
• Selected congress topic
• Keywords (min. 3 - max. 4)
• Bullets to summarize the paper and its highlights (max. 5)
• Abstract text (max. 2500 characters incl. spaces)
• Personal data of the co-author(s) (if any)

The results of the selection procedure will be sent by the 28 February 2014.

Events & Training Courses

• 7th International Conference on Computational and Financial Econometrics (CFE 2013)
14-16 December 2013, London
The Conference is organized by the London School of Economics, Queen Mary University of London and the Computational and Financial Econometrics (CFEnetwork). The CFEnetwork focuses on the interface of theoretical and applied econometrics, financial econometrics and computation in order to advance a powerful interdisciplinary research field with immediate applications. It aims to consolidate the research in CFE that is scattered throughout Europe, provide researchers with a network from which they can obtain an unrivalled source of information about the most recent developments in CFE as well as its applications, and to edit quality publications of high impact and significance. NAG is once again delighted to be supporting this event, and will be present at our exhibition booth.

Training Courses Provided by NAG's HECToR Team*

These HPC training courses are provided free of charge to HECToR users and UK academics whose work is covered by the remit of one of the participating research councils (EPSRC, NERC and BBSRC). The courses are also open to non-eligible people but will require payment of a course fee. Please see the eligibility page for more details.

The Best of the Blog

Implied Volatility using Python's Pandas Library

Blog snippet: Python has some nice packages such as numpy, scipy, and matplotlib for numerical computing and data visualization. Another package that deserves a mention that we have seen increasingly is Python's pandas library. Pandas has fast and efficient data analysis tools to store and process large amounts of data. Additionally, pandas has numpy and ctypes built into it which allow easy integration with NAG's nag4py package.

Read the full post by Brian Spector here.

NAGNews - Past Issues