Mathx::Stat and Bayes::Learn release 0.1.2
[email protected] (The Holy Ghost) Tue, 16 Oct 2018 09:17:51 +0200 (CEST)
| Newsgroups | perl.perl6.compiler |
|---|---|
| Message-ID | <[email protected]> |
------=_Part_384976_122533933.1539674271218
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Hello,
Mathx::Stat and Bayes::Learn should be indexable from holyghost's Perl6 CPAN directory.
Mathx::Stat is a statistics package unto correlation dimension. Variance, Estimate are distribution (population) based to enter more data which is processed. There are also Correlation and Covariance main pattern class functions. A probability class provides P(A and B), P(A or B), P(A|B) and so on.
Bayes::Learn has a minimax value function which can be used to calculate the infimum for least probability on it. There are also Bayesian Loss functions for distributions (based on Mathx::Stat distributions and population lists, which it's sole dependancy.)
Note that joint and disjoint experiments are available in the methods of both packages.
Both packages were written for speed in e.g. RTS or RPG games. So I didn't go the self dispatching way of OOP.
I will try to reach math people with Mathx which use Matlab or Mathematica for their work, but I will first make a markov strategy kit out of Mathx (not the Bayes package which is something different.) Later on there will be numerical methods (maybe not by me) needed for serious math work as I mentioned above.
My markov strategy kit will use Time chains, Markov Chains based on virtual nanoseconds also known as ticks. It will feature a non-extended Hidden Markov Model together with a Exponential distribution random number classification known as a truncated Gauss distribution.
I hope you all will enjoy this if you build Perl6 games out of SDL2::Raw for graphics.
All systems were built on Linux and might be adapted for BSD if I find the time.
Holly
------=_Part_384976_122533933.1539674271218
MIME-Version: 1.0
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
<!DOCTYPE html>
<html><head>
<meta charset=3D"UTF-8">
</head><body><p>Hello,</p><p><br></p><p>Mathx::Stat and Bayes::Learn should=
be indexable from holyghost's Perl6 CPAN directory.</p><p><br></p><p>M=
athx::Stat is a statistics package unto correlation dimension. Variance, Es=
timate are distribution (population) based to enter more data which is proc=
essed. There are also Correlation and Covariance main pattern class functio=
ns. A probability class provides P(A and B), P(A or B), P(A|B) and so on.</=
p><p>Bayes::Learn has a minimax value function which can be used to calcula=
te the infimum for least probability on it. There are also Bayesian Loss fu=
nctions for distributions (based on Mathx::Stat distributions and populatio=
n lists, which it's sole dependancy.)</p><p>Note that joint and disjoin=
t experiments are available in the methods of both packages.</p><p><br></p>=
<p>Both packages were written for speed in e.g. RTS or RPG games. So I didn=
't go the self dispatching way of OOP.</p><p><br></p><p>I will try to r=
each math people with Mathx which use Matlab or Mathematica for their work,=
but I will first make a markov strategy kit out of Mathx (not the Bayes pa=
ckage which is something different.) Later on there will be numerical metho=
ds (maybe not by me) needed for serious math work as I mentioned above.</p>=
<p><br></p><p>My markov strategy kit will use Time chains, Markov Chains ba=
sed on virtual nanoseconds also known as ticks. It will feature a non-exten=
ded Hidden Markov Model together with a Exponential distribution random num=
ber classification known as a truncated Gauss distribution.</p><p><br></p><=
p>I hope you all will enjoy this if you build Perl6 games out of SDL2::Raw =
for graphics.</p><p><br></p><p>All systems were built on Linux and might be=
adapted for BSD if I find the time.</p><p><br></p><p>Holly</p></body></htm=
l>
=20
------=_Part_384976_122533933.1539674271218--