The NAG C Library, Mark 6, is enhanced with nearly 60 new user-callable algorithms in statistics, sparse optimisation, polynomial root finding and special functions. In response to users' requests for more statistical algorithms, more than 30 new functions in ANOVA, nonparametric statistics, survival analysis and time series analysis (including GARCH) are included. These are in addition to the existing range of functions including ODEs optimisation, statistics and linear algebra.
The new nonparametric statistical algorithms enable scientists and engineers to make more robust decisions, as a result of making fewer assumptions in their data analysis.
GARCH algorithms are now available for the first time from NAG. These enable portfolio managers and others to model historical volatility and therefore to forecast volatility more accurately in the future.
The new nonlinear sparse optimiser is designed for people wishing to solve large-scale optimisation problems, such as those arising in the finance, engineering and aerospace industries. A sparse optimiser is much more appropriate when the models are large, since it takes advantage of the sparsity of the problem and therefore uses less memory and results in faster performance. In finance, this algorithm will enable portfolio managers to optimise portfolios that contain a large number of different assets, using more complicated models than the standard Markowitz model.
The NAG C Library is subjected to rigorous QA testing and only when it satisfies NAG's stringent accuracy standards is it released. C/C++ programmers will find it is easy to integrate the NAG C algorithms into existing applications and programmes such as C, C++ and Java. In addition, the thread-safe functionality allows them the flexibility to take advantage of developing multithreaded applications.
The C Library, which conforms to the ANSI C standard, will be initially available for Solaris Sun Sparc and Compaq Alpha Tru64. The implementations for Linux, HP, IBM, SGI and Windows will follow shortly.