Ryan Curtin
67b64d8a89
Loosen tolerance a bit, since it seems to fail once in a while. It's definitely
...
not broken though.
2014-12-03 02:17:33 +00:00
Ryan Curtin
0fa1b465bf
Fix convergence criterion according to Nishant's suggestion.
2014-12-03 01:59:28 +00:00
Ryan Curtin
f17f8b0cf2
Handle negative gradient values correctly.
2014-12-02 18:25:46 +00:00
Ryan Curtin
d8cbcdeaf1
Slightly widen tolerance.
2014-11-26 16:48:15 +00:00
Ryan Curtin
92508aa9ac
Accidentally checked in unstable code.
2014-11-26 16:47:57 +00:00
Ryan Curtin
ce04a249a7
Minor tolerance widening.
2014-11-26 16:16:12 +00:00
Ryan Curtin
6483b18605
Better handling of small gradient values.
2014-11-26 16:15:05 +00:00
Ryan Curtin
db174b1809
Slightly loosen tolerance.
2014-11-26 16:13:51 +00:00
Ryan Curtin
5504bb0d78
Tighten convergence tolerance for RastrigrinFunctionTest, since it doesn't seem
...
to be coming close enough to the desired minimum.
2014-11-25 16:57:36 +00:00
Ryan Curtin
4d014e856a
Reduce noise slightly and increase dataset size, which will slow down the test
...
but make the results more accurate.
2014-11-25 16:57:08 +00:00
Ryan Curtin
3e131eb8d6
Comment out NoCholeskySingularityTest in accordance with #373 .
2014-11-25 16:56:08 +00:00
Ryan Curtin
4471cac8df
Check frobenius norm overall instead of just for one element.
2014-11-25 16:22:37 +00:00
Ryan Curtin
7fdf21336f
Widen tolerances slightly. Maybe this test scheme isn't the best?
2014-11-24 23:07:08 +00:00
Ryan Curtin
5a1e5e74c2
Widen tolerance slightly.
2014-11-24 23:05:57 +00:00
Ryan Curtin
9bde6b4785
Widen tolerance slightly.
2014-11-24 23:05:46 +00:00
Ryan Curtin
2b381f4d35
Fix memory leak.
2014-11-24 23:05:16 +00:00
Ryan Curtin
dde478a53c
Significantly shrink size of test dataset because this test was taking 10
...
minutes.
2014-11-24 22:51:02 +00:00
Ryan Curtin
529a28152f
Handle setting seed properly for Armadillo RNGs past 3.930.
2014-11-19 18:52:57 +00:00
Ryan Curtin
00808b7fb5
Disable C4519 in prereqs.hpp not core.hpp.
2014-11-19 18:38:40 +00:00
Ryan Curtin
77ebf5159b
Include prereqs.hpp for compiler definitions and adjustments.
2014-11-19 18:38:26 +00:00
Ryan Curtin
ce6d5809e8
Disable C4519 errors entirely.
2014-11-19 18:32:46 +00:00
Ryan Curtin
f039262e46
Refactor test with negative elements to decompose the random matrix into its
...
proper low-rank decomposition, then test the reconstructed matrix.
2014-11-19 17:18:01 +00:00
Ryan Curtin
68ee9bae27
Remove debugging output.
2014-11-19 17:11:11 +00:00
Ryan Curtin
eb6333fa60
Somehow this never got added to the CMakeLists.txt.
2014-11-19 17:09:42 +00:00
Ryan Curtin
4b5d4ff1f4
Minor code cleanups.
2014-11-19 17:09:01 +00:00
Ryan Curtin
8872fb790b
Refactor for cleaner code and avoid storing WH explicitly if possible.
2014-11-19 17:08:48 +00:00
Ryan Curtin
0ccb8e3d22
Refactor GeneralizedRosenbrockTest to deal with intermittent failures better.
...
Also use 4 trials for RastigrinFunctionTest.
2014-11-18 23:20:31 +00:00
Ryan Curtin
fed7bf245c
Issue a runtime error if the user is using Armadillo 4.300.x through 4.400.x and
...
tries to load or save HDF5 files, since that is a bug in Armadillo.
2014-11-18 22:55:49 +00:00
Ryan Curtin
04d4d3d054
Don't test on Armadillo 4.300.0 through 4.400.x because there is a bug in
...
Mat::load(istream&) which prevents loading from type hdf5_binary. (The bug is
simply the omission of the hdf5_binary case from the switch() statement
actually.)
2014-11-18 22:55:23 +00:00
Ryan Curtin
a63e408ee6
Fix uninitialized memory issue (dsPredictions was never set).
2014-11-18 22:03:12 +00:00
Ryan Curtin
b0ba1b4ff5
Nope, turns out I am wrong. C++03, C++11, and C++14 all assume a program
...
reaching the end of main() without returning anything will return 0.
2014-11-18 21:53:18 +00:00
Ryan Curtin
0f2aa34581
Pedanticism: return a value at the end of main().
2014-11-18 21:36:23 +00:00
Ryan Curtin
6d60e7c14f
Fix -Wunintialized, reported by govg.
2014-11-18 20:46:28 +00:00
Ryan Curtin
b7cbda9d96
Slightly loosen tolerances for NMF tests.
2014-11-17 18:50:27 +00:00
michaelfox99
5159126f17
Re-ordered initializer lists to fix warnings
2014-11-15 22:18:25 +00:00
michaelfox99
b9444785e4
Dimensionality() now returns proper values
2014-11-15 19:40:47 +00:00
michaelfox99
7020105c7f
hmm_regression.hpp and hmm_regression_impl.hpp added
2014-11-15 19:35:34 +00:00
michaelfox99
5011d665a2
Implementation of HMMRegression class
2014-11-15 19:34:22 +00:00
michaelfox99
52a43055e8
HMM regression method
2014-11-15 19:33:59 +00:00
michaelfox99
d840852ca0
transition now protected, not private
2014-11-15 19:32:30 +00:00
Ryan Curtin
230bd4fa69
Better handling of the weird case when includes are needed but the library isn't.
2014-11-14 16:44:40 +00:00
Ryan Curtin
e4a0ff4c5b
Refactor CountMostFreq() so it is faster, simpler, and doesn't sometimes return
...
uninitialized values.
2014-11-11 18:46:31 +00:00
Ryan Curtin
695e59168f
The calculation here was actually incorrect.
2014-11-11 18:46:00 +00:00
Ryan Curtin
e45ce79e2a
Smarter handling of HDF5 dependency search, especially for Debian systems where
...
things are Weird(TM).
2014-11-11 17:15:11 +00:00
Ryan Curtin
631955eb4c
Not sure how I missed this spelling error...
2014-11-10 19:12:48 +00:00
Ryan Curtin
4ff522b053
If gradient2 or gradient1 are zero, then BOOST_REQUIRE_CLOSE will fail, so use
...
BOOST_REQUIRE_SMALL in those situations.
2014-11-10 16:56:39 +00:00
Ryan Curtin
8da5cc01b1
strlen() returns the length of the string but you must account for the null
...
terminator yourself. Hence, this code sometimes caused random invalid writes
and crashes.
2014-11-10 16:54:12 +00:00
Ryan Curtin
783f28bcf5
Remove debugging random seed.
2014-11-10 16:47:00 +00:00
Ryan Curtin
4a673018b4
Loosen tolerance until a better solution is devised (currently I am waiting on
...
an email from Nishant).
2014-11-10 16:45:10 +00:00
Ryan Curtin
81a6e93784
Add Pelleg-Moore type prune. This improves performance -- at least a bit.
2014-11-10 16:33:51 +00:00
Ryan Curtin