NAGnews 133

Posted on
26 Aug 2015

In this issue:

Using Principal Component Analysis to model Yield Curves

NAG and Teggin Consulting have co-authored a paper which applies Principal Component Analysis (PCA) to yield curve data. Modelling all the points on the yield curve is highly complex, so actuaries have traditionally focused on modelling parallel shifts to the yield curve. This is straightforward and captures much of the impact of yield curve movements. However, pressure for economic realism, both from firms and from regulators, is increasingly driving a need for richer approaches.

PCA is a technique which allows focus to be given to the most important elements of variation in a high-dimensional data set such as yield curve data. We shall see that, with only a small degree of approximation, the complexity of modelling all the points on the yield curve can be reduced to modelling a much smaller number of key components.

Click here to read the white paper.

NAG helps to approve the W3C MathML 3.0 ISO/IEC International Standard

For many years NAG has supported the development of the W3C MathML standard. David Carlisle, Principal Technical Consultant at NAG, is currently, with Patrick Ion of the American Mathematical Society, co-chair of the W3C Math Working Group and editor of the MathML Specification.

In June 2015, the Joint Technical Committee JTC 1 of the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC), announced approval of the MathML Version 3.0 2nd Edition as an ISO/IEC International Standard (ISO/IEC 40314:2015).

This ISO standard version of MathML is identical in technical content to the W3C MathML 3 2nd Edition Recommendation, but formalises the format as an international standard.

NAG uses MathML for the mathematical content in all its documentation and uses it internally for code generation and other projects. We welcome its adoption as an ISO/IEC standard.

Learn more here.

Mark 25 news/spotlight: Change Point Analysis

New numerical routines for performing Change Point Analysis now feature in the latest mark of the NAG Library as a direct result of code contribution from Dr Rebecca Killick, University of Leicester. Dr Killick donated Change Point Analysis code which was then put through the "NAG Engine", tested, documented and made available for users of the Library.

Change Point Analysis Background

Given a series of data, change point analysis involves detecting the number and location of change points, locations in the data where some feature, for example the mean, changes. This has applications in a wide variety of areas including: finance, quality control, genetics, environmental studies and medicine.

Read more about this exciting new NAG Library functionality here.

ARM to leverage NAG's numerical engineering expertise in its new ARM Math Libraries

At the recent ISC HPC conference in Frankfurt, ARM and NAG announced plans to offer a commercially supported set of 64-bit ARMv8 numerical libraries for scientific computing. ARM will leverage NAG's expert numerical engineering capabilities to create a unified framework and validation suite.

arm's notification

Best of the Blog

NAG Linear Regression on Apache Spark

In this post we test the scalability and performance of using NAG Library for Java to solve a large-scale multi-linear regression problem on Spark. We solve this problem using the normal equations. In the final step, a NAG linear regression routine is called on the master node to calculate the regression coefficients. All of this happens in one pass over the data - no iterative methods needed! Read more.

Training Courses & Events

NAG will be at the following exhibitions and conferences over the next few months.

NAGnews - Past Issues

We provide an online archive of past issues of NAGnews. For editions prior to 2010, please contact us.