aftermath Wiki Rss Feedhttps://aftermath.codeplex.com/aftermath Wiki Rss DescriptionUpdated Wiki: Homehttps://aftermath.codeplex.com/wikipage?version=7<div class="wikidoc"><b>Aftermath</b><br />Set of C++11 header files providing support for---mainly probabilistic and statistical---mathematical tasks, such as pseudorandom number generation.<br /><br />Currently contains the following sections:
<ul><li><b>algebra</b> Some generic math that might or might not be useful at times.</li>
<li><b>numerics</b> Provides sse/avx implementation for matrices and simple mathematical operations, such as addition / multiplication.</li>
<li><b>probability</b> Contains some basic distribution types; as opposed to standard c++, the intention is to provide a framework for working with mathematical/statistical properties of said distributions; (pseudo-) random number generation is dealt with in the next section (<i>random</i>).</li>
<li><b>random</b> (<i>depends on probability</i>) Provides sampling methods from some probability distributions. It features an implementation of ziggurat method (see <i>Marsaglia and Tsang [undated]</i> and <i>Doornik [1997]</i>) for generating normal and exponential random variables, offering a significant speed-up compared to the default c++ implementation.</li></ul>
<br /><i>Note on the deviation from standard c++ distribution/sampling</i>. The current standard is to let distribution types perform sampling via <span class="codeInline">operator ()</span>; we, on the other hand, introduce <i>sampler types</i> to perform the sampling via the same <span class="codeInline">operator ()</span> interface, and let <i>distribution types</i> expose just the mathematical properties of the distributions. We think this is an important distinction; it also covers the standard case in that c++ distribution types are also sampler types.</div><div class="ClearBoth"></div>GregorySat, 25 Jul 2015 13:10:32 GMTUpdated Wiki: Home 20150725011032P