Database Reference
In-Depth Information
library, developed by the European Organization for Nuclear Research
(CERN, which is also where the web was invented). The library is licensed
using the GNU Lesser General Public License (LGPL), with an explicit
addendum prohibiting use in military applications. The Maven dependency
for the library is as follows:
<dependency>
<groupId>colt</groupId>
<artifactId>colt</artifactId>
<version>1.2.0</version>
</dependency>
The Apache Commons libraries also provide Java libraries for numerical
computing.
Other languages also have good numerical libraries, often providing
interfaces to high-performance and well-tested Fortran libraries. For
commercial applications, the NAG Library and International Mathematics
and Statistics Library (IMSL) have long been popular as they have libraries
available for a variety of languages.
Users of the C language can use open source alternatives such as the GNU
Scientific Library (
http://www.gnu.org/software/gsl/
)
as an alternative to
the Colt libraries used in this topic. Another alternative is the embedded
version of the R Statistical Language (
http://r-project.org
), which makes
available many of its core numerical routines. For C++, the Boost library
(
http://boost .org
)
is a popular choice, powering a number of applications.
For high-level languages, the options are more sparse. Python probably
has the most developed numerical libraries in the form of NumPy
(
http://numpy.org
)
and SciPy (
http://scipy.org
)
. Other languages, such as
Ruby and JavaScript, have special-purpose libraries available, but nothing
that has really been collected into a comprehensive numerical computing
library.
Probabilities and Distributions
Probability theory underpins the entire field of statistics. Originally
developed to understand games and gambling, the classical application is
the study of one or more urns full of colored balls: “What is the probability
that the next ball drawn from the urn will be red?” Of course, the answer to
