C++ Normal Distribution

Question

Possible Duplicate:
erf(x) and math.h
Best library for statistics in C++?

I am wondering if there is any standard way to implement a cumulative normal distribution function in C++. Or, is there any existing library for that task?

Can anyone recommend a good statistics library for C++, if there is any?

Thanks.

If you would be more specific about what kind of functionality you expect that library to have we could probably give better recommendations. Boost provides some basic tools, and then there are libraries like gsl or ROOT which go further. But it doesn't end there. — Benjamin Bannier, Sep 28 '12 at 17:37

CrazyCasta · Accepted Answer · 2012-09-28T17:46:29.027

4

You can do this with the erf function (which is include in the standard math library, source).

CDF = 0.5 + 0.5 * erf(x/(sigma * sqrt(2.0)));

sigma = 1.0 for a normal distribution of course.

edited Sep 28 '12 at 17:46

answered Sep 28 '12 at 17:37

CrazyCasta

26,917
4
45
72

Ahh. Beat me to it. You don't need boost. All you need is ``. Plus 1. – David Hammen Sep 28 '12 at 18:16

score 1 · Answer 2 · answered Sep 28 '12 at 17:43

1

Boost is as good as the standard. For more boost maths/statistical.

Simple example for stand-alone C++ implementation of the cumulative normal distribution is available Here

answered Sep 28 '12 at 17:43

nKandel

2,543
1
29
47

1

You don't need Boost. Not really. The relation between `erf()` and `phi()` is a trivial one. – David Hammen Sep 28 '12 at 18:17
1

He is asking for libraries and Boost is one choice. – Benjamin Bannier Sep 28 '12 at 18:19
And the C math library is another. Why go to boost when this is already native? – David Hammen Sep 28 '12 at 18:21
@DavidHammen: `libm` isn't exactly famous for the many statistical tools it provides. But again, the question isn't well put. – Benjamin Bannier Sep 28 '12 at 18:26

C++ Normal Distribution

2 Answers2