2
0
mirror of https://github.com/boostorg/math.git synced 2026-01-30 08:02:11 +00:00
Commit Graph

199 Commits

Author SHA1 Message Date
John Maddock
1002d7d090 Minor coments/changes to code.
Proof reading changes to docs, mostly rewritten neg binomial worked example.
Removed dead test.


[SVN r3473]
2006-11-28 14:12:26 +00:00
John Maddock
475d4a0280 Misc small fixes and warning suppressions.
[SVN r3472]
2006-11-28 10:54:09 +00:00
Paul A. Bristow
69ecd8d2a0 Boundary values for quantile corrected and tests added but still don't deal with infinity well. TODO still.
[SVN r3469]
2006-11-27 19:02:37 +00:00
Paul A. Bristow
2fdd390169 Promotion no longer relevant so removed..
[SVN r3468]
2006-11-27 15:40:28 +00:00
John Maddock
ed1b8ae751 Finally got the parameter estimators right, and checked against "known goods".
Misc small fixes to the docs and example programs.


[SVN r3462]
2006-11-26 09:03:49 +00:00
John Maddock
7fc815029e Fixed up neg binomial estimators.
Enabled neg binomial tests and adjusted arguments.
Added extra comment to binomial.


[SVN r3460]
2006-11-24 19:14:32 +00:00
John Maddock
d15bf712e5 Fixed binomial estimators and added stronger tests.
Typos in concepts.
Enabled native log1pl on more platforms.
Fixed binomial example.
Added image.


[SVN r3459]
2006-11-24 18:54:52 +00:00
Paul A. Bristow
867dd1d8e6 added code for special case prob < cdf(0) for quantile and complement.
[SVN r3458]
2006-11-24 17:53:10 +00:00
Paul A. Bristow
4a6cd24f9c suppressed error 4702 unreachable code.
[SVN r3451]
2006-11-24 12:23:32 +00:00
Paul A. Bristow
48b5f13838 if (p <= pow(1 - dist.success_fraction(), dist.trials()))
{ // p <= pdf(dist, 0) == cdf(dist, 0)
   return 0; // So the only reasonable result is zero.
} // And root finder would fail otherwise.
Added, and similarly for complement


[SVN r3445]
2006-11-23 17:46:15 +00:00
John Maddock
78244650dc More fixes to precision.hpp
More or less finished concept docs.
Lots of proof-reading and editorial fixes.


[SVN r3440]
2006-11-22 18:48:28 +00:00
Paul A. Bristow
24af492a2d Correction to estimate parameters.
[SVN r3437]
2006-11-22 18:28:23 +00:00
Paul A. Bristow
1192252ba9 Correced typos and added <16834> size for 80 and 128 bit
[SVN r3436]
2006-11-22 16:48:52 +00:00
John Maddock
12984831f0 More concept docs, updated code to match.
[SVN r3423]
2006-11-21 13:19:08 +00:00
John Maddock
b89f35783e More editorial and proof reading fixes.
Added initial draft of Remez background information.


[SVN r3414]
2006-11-19 18:29:06 +00:00
Paul A. Bristow
a6b95e985f Some corrections but still not right.
[SVN r3409]
2006-11-16 12:46:45 +00:00
Paul A. Bristow
7cd4885c79 Nano optimisation of skewness.
[SVN r3398]
2006-11-15 12:43:32 +00:00
Paul A. Bristow
3242af54a7 Minor revisions
[SVN r3389]
2006-11-14 15:40:12 +00:00
Paul A. Bristow
8452c720ba conditional expression is constant warning 4127 suppressed. (produced 4 pages of messages!)
[SVN r3384]
2006-11-14 11:48:19 +00:00
Paul A. Bristow
f84fef66c8 More tests including test_spot and handful of spot values.
[SVN r3362]
2006-11-09 17:33:17 +00:00
John Maddock
3ea5308b45 Fixed problem with chi-squared sample size estimator.
Updated test program accordingly.
Added chi-square tutorial code.


[SVN r3360]
2006-11-08 17:14:39 +00:00
Paul A. Bristow
32be9aa3f3 corrected quantile to use ibeta_inv. test_spot still todo. & parameter estimate??
[SVN r3358]
2006-11-08 15:27:01 +00:00
Paul A. Bristow
a4c6015e0b 1st cut with a few simple tests for double only.
[SVN r3354]
2006-11-07 12:21:22 +00:00
Paul A. Bristow
24d36b78f6 Spelling correction, and comments copied to concepts.qbk.
[SVN r3352]
2006-11-07 09:35:58 +00:00
John Maddock
9e41ae39b4 Minor fixes for gcc in poisson.hpp
Tidied up fraction.hpp.
Updated linux acceptable error rates.


[SVN r3346]
2006-11-04 18:28:24 +00:00
Paul A. Bristow
d9d5f2b21e tests revised to try to cover failures in 80 and 128 bit FP cases. Added throw checks
[SVN r3345]
2006-11-04 10:55:12 +00:00
John Maddock
feacc055d3 Added gamma distribution and tests.
Fixed a lot of gcc-linux issues: mostly missing "using namespace std" declarations.
Added new concept check to test for these kinds of failures at compile time.


[SVN r3344]
2006-11-03 18:42:23 +00:00
Paul A. Bristow
041d476768 Removed median. Allowed hazard & chf to be provided by derived accessors, and removed any mention of std::numeric_limits<T>::quiet_NaN()
[SVN r3343]
2006-11-03 15:04:38 +00:00
John Maddock
8441e358fb Rewritten tgamma1pm1 in terms of lgamma rational approximations: now more efficient and more accurate!
Minor patch to exponential distribution tests.
Removed dead code from Remez app.


[SVN r3331]
2006-10-31 10:50:11 +00:00
John Maddock
7d003a280f Fix typo and regressions.
[SVN r3330]
2006-10-31 10:48:26 +00:00
John Maddock
2132f92495 Changed tgamma to use factorial lookup where applicable.
Updated docs accordingly and added new docs for the new factorial and binomial functions.


[SVN r3327]
2006-10-28 16:30:39 +00:00
John Maddock
bcbb6f3ca9 Added double_factorial plus rising and falling factorials and binomial coefficients.
Added new tests for the above.
Refactored factorials header.
Optimised tgamma_delta_ratio for integer arguments.


[SVN r3326]
2006-10-28 13:00:10 +00:00
Paul A. Bristow
da6a644de8 Minor comments edits, for more tests OK with builint float double but fail neg binomial for real concept.
[SVN r3324]
2006-10-26 13:12:20 +00:00
John Maddock
82548699a5 Improved and generally tidied up sqrt1pm1 and powm1.
Renamed sqrtp1m1 as sqrt1pm1 to match log1p etc.
Added docs for misc power functions, plus weibull distribution.
Updated tests to match the above changes.


[SVN r3321]
2006-10-26 11:51:59 +00:00
John Maddock
527bf0d227 Updated erf implementation selection logic.
Provided rational expansions for expm1.
Provided faster form for log1p.
Updated remez minimax code.


[SVN r3319]
2006-10-25 18:10:13 +00:00
John Maddock
9b16307ae7 Fixed a bug in incomplete gamma, and refactored and refined internal selection logic.
Added tests for incomplete gamma for int and half int arguments.


[SVN r3312]
2006-10-25 10:19:34 +00:00
John Maddock
24066f892f Updated to call overflow_error when the result is infinite.
[SVN r3311]
2006-10-25 09:38:24 +00:00
John Maddock
2f964f6ef6 Added finite sum to incomplete beta, plus tests and docs.
Linked in lognormal docs.


[SVN r3310]
2006-10-25 09:03:04 +00:00
Paul A. Bristow
76f2122bad Comments.
[SVN r3309]
2006-10-24 18:05:48 +00:00
Paul A. Bristow
84bdaaa6ee Comments and checks that no return of infinity that should call overflow_error.
[SVN r3308]
2006-10-24 17:58:30 +00:00
Paul A. Bristow
25570182cb Tests now pass, but more tests to do still.
[SVN r3307]
2006-10-24 17:47:40 +00:00
Paul A. Bristow
504e9da847 Layout & comments only.
[SVN r3303]
2006-10-24 09:36:36 +00:00
John Maddock
7a0ac09ba8 Misc fixes and better error handling and special case treatment.
[SVN r3300]
2006-10-23 12:55:19 +00:00
Paul A. Bristow
0a4dfe1bc5 complement and derived accessors still TODO
[SVN r3299]
2006-10-23 12:01:03 +00:00
Paul A. Bristow
0af1942639 Added complement.hpp and expm1 to get it compile, but fails tests as yet.
[SVN r3297]
2006-10-23 11:39:43 +00:00
Paul A. Bristow
d29c25757c Added some checks for scale and prob. But fails tests.
[SVN r3292]
2006-10-23 10:09:57 +00:00
Paul A. Bristow
c7c1e86552 Several "return static_cast<V>( ... )" added to avoid warning C4244 conversion from 'const double' to 'float'.
Not all tested by test_chi_squared usage :-(  so a mistake might cause compile failures elsewhere?


[SVN r3283]
2006-10-22 16:05:07 +00:00
Paul A. Bristow
ecd6a66fee Added default RealType double (except binomial - where commented out).
(Changes to test in binomial reflect this, others in progress).


[SVN r3278]
2006-10-21 18:34:13 +00:00
Paul A. Bristow
860e361ef7 default Real type double
[SVN r3277]
2006-10-21 18:07:46 +00:00
John Maddock
b6529add8a Added preliminary lognormal and weibull implementations.
Updated common error handling.
Added more test values.


[SVN r3265]
2006-10-18 15:18:15 +00:00