boost/math
2
0
Fork 0
mirror of https://github.com/boostorg/math.git synced 2026-01-28 19:32:08 +00:00
Code Issues Projects Releases Wiki Activity
2,114 Commits 37 Branches 125 Tags
fc112cd24fac8474ce005ec701300241d4e65a0a
Commit Graph

835 Commits

Author SHA1 Message Date
John Maddock
fc112cd24f Don't pass potential expression templates to make_tuple as it confuses gcc-4.4.x. 
[SVN r81546]
 2012-11-26 10:23:17 +00:00
John Maddock
3822a35591 Minor fixes/typo corrections. 
[SVN r81424]
 2012-11-19 13:35:49 +00:00
John Maddock
d6320c49c0 Minor fix for expression template types. 
[SVN r81120]
 2012-10-31 17:04:47 +00:00
John Maddock
ba13586027 Fix asinh and acosh to work with very large arguments without overflow. 
[SVN r80884]
 2012-10-06 16:53:15 +00:00
John Maddock
b9e891404a Fix improper overflow error handling in asinh.hpp. 
Update asinh and acosh test cases.
Add fixes to release notes and regenerate docs.
Fixes #7416.

[SVN r80695]
 2012-09-24 17:16:29 +00:00
John Maddock
068e88fa36 Fix atanh error handling. 
Add extra tests to catch above bug.
Fixes #7415.

[SVN r80692]
 2012-09-24 16:23:06 +00:00
Paul A. Bristow
a286b00833 delta == 0 treated as special case for mean, variance, skewness and kurtosis 
== central Student's t

[SVN r80233]
 2012-08-26 09:43:43 +00:00
Paul A. Bristow
4a1ccd574e Comment only on change to use normal for 
(v > 1 / boost::math::tools::epsilon<T>() 

[SVN r80196]
 2012-08-25 10:13:27 +00:00
John Maddock
e5723529c2 Fix remaining GCC and Intel issues for very large v in non-central t. 
[SVN r80195]
 2012-08-25 09:13:42 +00:00
Paul A. Bristow
c1b086325d Use normal distribution for v > 1/eps 
[SVN r80156]
 2012-08-23 13:43:44 +00:00
Paul A. Bristow
ef6a5b791a Added support for infinite degrees of freedom. 
Correct some error messages (and made quantile produce the correct complemented version by forwarding the function string to the detail:: code).

There  are still some confusing variable names.



[SVN r80118]
 2012-08-21 17:39:51 +00:00
Paul A. Bristow
1b893e541b Added df >0 and infinite test, used by students t and non-central t. 
[SVN r80117]
 2012-08-21 17:37:20 +00:00
John Maddock
885e35a8fd Update reference in comments. 
[SVN r80070]
 2012-08-17 16:19:23 +00:00
John Maddock
9cef5e424d Change variable names to match common usage. 
[SVN r80069]
 2012-08-17 16:04:15 +00:00
John Maddock
1f9cdaf16f Add needed using declarations. 
[SVN r80068]
 2012-08-17 15:32:12 +00:00
John Maddock
29d598eb97 Add student's t approximations for large v. 
[SVN r80066]
 2012-08-17 12:21:57 +00:00
John Maddock
4be8acfe68 Fix numerical stability problem in incomplete beta: ibeta_fraction2 sometimes has catastrophic cancellation errors when a and b are large and x is near 1. 
Change non central t series termination conditions to check for divergent series.
Update tests with new larger test values.

[SVN r80065]
 2012-08-17 11:18:18 +00:00
John Maddock
bf39b3834a Fix bug in incomplete beta which causes an exception to be thrown in itrunc when one argument is very large. 
Fix several similar issue in the non-central T which cause cancellation errors if we let the starting iteration be zero.
Remove dead code that results from above changes.
Minor fixes to table_type and test_next.cpp.

[SVN r80061]
 2012-08-16 15:27:40 +00:00
John Maddock
a4640b3cf4 Fix typo in algorithm selection. 
[SVN r79982]
 2012-08-12 17:03:26 +00:00
Paul A. Bristow
e4cf24554b Changes to deal with copyright and tabs from inspection report 
[SVN r79949]
 2012-08-09 17:58:46 +00:00
Paul A. Bristow
cf1d644c4e Major update to allow df == +infinity. 
[SVN r79911]
 2012-08-07 15:53:54 +00:00
Paul A. Bristow
5d887432cd Using the 1/eps to switch to normal distribution. 
[SVN r79892]
 2012-08-06 16:30:12 +00:00
John Maddock
5ed9615fb5 Fixes for multiprecision and expression template enabled types. 
[SVN r79851]
 2012-08-02 15:16:21 +00:00
Paul A. Bristow
139a12844e Back to plan A! not defined for v <= k - no infinity from moments. 
(TODO use normal for df = infinity?)

[SVN r79836]
 2012-08-01 17:11:31 +00:00
Paul A. Bristow
b3bee00ce0 Various cosmetic changes to deal with Inspection nags. 
[SVN r79818]
 2012-07-31 13:32:25 +00:00
Paul A. Bristow
059d1b4378 Added terminal newline. 
[SVN r79814]
 2012-07-31 11:28:17 +00:00
Paul A. Bristow
d7dc107b30 Remove tabs and () max call. 
[SVN r79813]
 2012-07-31 11:26:45 +00:00
Paul A. Bristow
848c0da5c9 Removed tabs 
[SVN r79811]
 2012-07-31 11:15:30 +00:00
Paul A. Bristow
4ad388a10e Remove non-ASCII char in comment. 
[SVN r79810]
 2012-07-31 10:57:22 +00:00
Paul A. Bristow
e71253c782 Revised checks on moments and return values 
[SVN r79770]
 2012-07-27 15:34:35 +00:00
Paul A. Bristow
b12334ba6e Revised for Trac #7717 
[SVN r79759]
 2012-07-26 18:03:32 +00:00
John Maddock
c2e90167f7 Tentative fix for issue #4445. 
Adds tests for next.hpp when using the SSE2 instruction set in DAZ and FTZ modes.
Note that FTZ mode is particularly pernicious as non-zeros can change to zeros at random points in program flow depending upon the exact instructions generated by the compiler - thus breaking program logic/invariants.
Refs #4445.

[SVN r79572]
 2012-07-17 09:22:59 +00:00
John Maddock
c3da72b1b2 Fix typo in Hankel function implementation that causes incorrect values to be returned for x < 0 and v an odd integer. 
Add additional Hankel function tests.
Add a couple of missing header inclusion tests.
Update history in docs, and regenerate HTML.
Fixes #7135.

[SVN r79510]
 2012-07-14 18:34:01 +00:00
John Maddock
cfd28e0e05 Tentative fix #2 for TI compiler error. 
Refs #7099.

[SVN r79439]
 2012-07-12 09:53:18 +00:00
John Maddock
c585419ec2 Allow fpclassify and friends to be used with integer types via the usual promotion mechanism. 
Fixes #6517.

[SVN r79438]
 2012-07-12 09:51:32 +00:00
John Maddock
58e250eb1f Ditch anonymous union - it causes GCC-4.4 to choke. 
Fixes #6362.

[SVN r79421]
 2012-07-11 17:28:54 +00:00
John Maddock
176a695ec0 Fix double inclusion of <complex> which causes use of Boost.TR1 to fail. 
Fixes #7053.

[SVN r79420]
 2012-07-11 17:27:27 +00:00
Marshall Clow
985d3cdc00 Switch from deprecated macros to new shiny ones; no functionality change 
[SVN r79417]
 2012-07-11 16:21:28 +00:00
John Maddock
f73f66b454 Tentative fix for compilers that treat floating point constants smaller than a 64-bit double can hold as hard errors. 
Refs #7099.

[SVN r79332]
 2012-07-07 11:18:09 +00:00
John Maddock
629db84174 Fix file line ending. 
Fixes #7062.
Fixes #7097.

[SVN r79317]
 2012-07-06 16:17:39 +00:00
John Maddock
19f78ac0ac Fix actual and potential integer divisions when using integer constants. 
Fixes #7024.

[SVN r79154]
 2012-06-28 12:09:09 +00:00
John Maddock
26d92a55f0 Add missing include guards. 
Fixes #7001.

[SVN r79027]
 2012-06-21 09:18:47 +00:00
John Maddock
6007215c95 Update more distribution tests for fisher_f and extreme_value and fix failures. 
Refs #6934.

[SVN r78783]
 2012-05-31 11:15:42 +00:00
Paul A. Bristow
3f93aff61d Added checks for non finite parameter and x = NaN 
[SVN r78781]
 2012-05-31 11:10:50 +00:00
John Maddock
73c9b11179 Update more distribution tests, from Inverse* down to gamma. 
Refs #6934.

[SVN r78778]
 2012-05-31 10:18:02 +00:00
Paul A. Bristow
4dd9452c22 Corrected failure to check for NaN in quantiles. UNsure if df should include zero. 
[SVN r78776]
 2012-05-31 09:35:52 +00:00
John Maddock
392fbeeddd Add extra error tests for negative_binomial to laplace distros. 
Refs #6934.

[SVN r78773]
 2012-05-30 18:15:41 +00:00
John Maddock
41bd4bd442 Update to use new out-of-bounds error checking. 
Refs #6934.

[SVN r78771]
 2012-05-30 16:52:18 +00:00
Paul A. Bristow
a72f149301 Changed range to infinity if supported. 
[SVN r78768]
 2012-05-30 16:35:59 +00:00
Paul A. Bristow
7ef962fc6b Changed range to include infinity if supported. 
[SVN r78766]
 2012-05-30 16:31:18 +00:00