Fixed handling of numeric variables in factorDesign - #47

Open

coreybrier wants to merge 455 commits into

selective-inference:masterfrom

coreybrier:corey

This pull request is big! We're only showing the most recent 250 commits

Commits on Aug 16, 2017

removed print statement
jonathan-taylor
committed
e87d12d
using C for each row
jonathan-taylor
committed
210d0fe
optimized the C code a bit -- still has debug statements
jonathan-taylor
committed
c7c41e4
removing debug statements, making objective computation faster
jonathan-taylor
committed
f6ae738
Merge pull request #18 from jonathan-taylor/master
tibshirani
authored
2bd33e8
slightly lower tolerance
jonathan-taylor
committed
1a8e2ad
using active sets
jonathan-taylor
committed
c391bd3

Commits on Aug 17, 2017

begun the Rcpp extension
jonathan-taylor
committed
096b0b8
trying to save some copy time -- R doesn't seem to care about extra arguments to .C
jonathan-taylor
committed
faf0436
don't lose this
jonathan-taylor
committed
5bf8018
segfaulting on the test_big.R :(
jonathan-taylor
committed
944ae7d
changes to funs.fixed.R that will be overwritten
jonathan-taylor
committed
4caf225
success -- about 2ms now -- python still faster (as timed through rpy2) :)
jonathan-taylor
committed
f53af23
BF: column major, removing print statements
jonathan-taylor
committed
7920c8c
changing name of function, evaluation of feasibility in Rcpp
jonathan-taylor
committed
6756ab2
RF: renamed C function, removing R solver for one row
jonathan-taylor
committed
0e58435
need to make Rcpp exports
jonathan-taylor
committed
0614305
ignore first two lines if status not OK?
jonathan-taylor
committed
ca60bc9
install Rcpp earlier
jonathan-taylor
committed
60f6cde
using Rcpp for matrixcomps as well -- update1 and downdate1 -- examples now OK on Rcmd check
jonathan-taylor
committed
c1aef63
using a warm start if available
jonathan-taylor
committed
f468510

Commits on Aug 18, 2017

trying to find segfault
jonathan-taylor
committed
a8805b8
BF: order of arguments to is_active, renamed the function as well
jonathan-taylor
committed
abf0351
removing debug statements
jonathan-taylor
committed
27af0d5
not using the C code
jonathan-taylor
committed
3a46ad8
removing useC
jonathan-taylor
committed
de103f7
BF: percent signs in example
jonathan-taylor
committed
d487526
smaller changes in mu per step in line search
jonathan-taylor
committed
7360386
renaming Sigma_theta to gradient, making sure we loop through active set
jonathan-taylor
committed
a9a8bbb
more renaming
jonathan-taylor
committed
7f1e4a6
BF: forgot to rename in Rcpp file, passing active and ever_active from R
jonathan-taylor
committed
a4055ce
more renaming
jonathan-taylor
committed
f432b42
trying to get arbitrary linfunc
jonathan-taylor
committed
fcd8206
can solve arbitrary LASSO now
jonathan-taylor
committed
a67d29f
BF: updating active set
jonathan-taylor
committed
d248a8f
cosmetic change
jonathan-taylor
committed
aa515c9
BF: fixing function signatures
jonathan-taylor
committed
5d5e8f3
BF: fixing function signatures
jonathan-taylor
committed
d1debcd
BF: KKT wrong for active conditions, still have debug statements
jonathan-taylor
committed
15762af
removing debug statements
jonathan-taylor
committed
9872b45
BF: KKT condition was wrong
jonathan-taylor
committed
be84df2
lenient tolerance
jonathan-taylor
committed
0de59af
added a few targets to makefile
jonathan-taylor
committed
190c0a0
not checking objective
jonathan-taylor
committed
2c4cf3a

Commits on Aug 19, 2017

Commits on Aug 20, 2017

Commits on Aug 21, 2017

Merge pull request #26 from jonathan-taylor/using_Rcpp
jonathan-taylor
authored
6dc80be

Commits on Aug 29, 2017

merging with upstream/master
jonathan-taylor
committed
3383e31
NF: stop early if active set gets too big
jonathan-taylor
committed
7fa5120

Commits on Aug 30, 2017

renaming function, but all linesearch logic into a row-wise function
jonathan-taylor
committed
0c1f9ea
camel case function name (ew... but consistent)
jonathan-taylor
committed
e292429
new export debiasingMatrix
jonathan-taylor
committed
ec07564
minor edit
jonathan-taylor
committed
3c0ef8d
BF: latex percent
jonathan-taylor
committed
d94d217
a little more doc
jonathan-taylor
committed
3e9d62d
Merge pull request #27 from jonathan-taylor/max_active
jonathan-taylor
authored
bbf7e19
merging with upstream master
jonathan-taylor
committed
3c3d868

Commits on Aug 31, 2017

exporting function to construct polyhedral constraints of LASSO
jonathan-taylor
committed
3d8b248
removing unused C code
jonathan-taylor
committed
d935d94
moving file back
jonathan-taylor
committed
7004a0b

Commits on Sep 1, 2017

Rcpp exports
jonathan-taylor
committed
5e671a4

Commits on Sep 11, 2017

Commits on Sep 22, 2017

copy of CRAN version 1.2.3 in selectiveInference-currentCRAN
jonathan-taylor
committed
7040ad4
new versions of RcppExports
jonathan-taylor
committed
0a3bc8d

Commits on Sep 25, 2017

Merge pull request #28 from jonathan-taylor/export_lasso_poly
jonathan-taylor
authored
5a0884b

Commits on Sep 26, 2017

fixing examples for new glmnet, small cleanup of function signature
jonathan-taylor
committed
59e9c7e
removing a tab
jonathan-taylor
committed
41dc63a
WIP: writing a solver for wide X matrices
jonathan-taylor
committed
86a92fd
for devel version -- needs intervals reinstalled
jonathan-taylor
committed
2f05bf6
Merge pull request #30 from jonathan-taylor/master
jonathan-taylor
authored
48547ea
WIP: working on wide X QP
jonathan-taylor
committed
6238a92
WIP: builds now, but segfaulting
jonathan-taylor
committed
dd30abc
test script for wide problem
jonathan-taylor
committed
ed19c29
test running now -- no segfault?
jonathan-taylor
committed
ce0dc07

Commits on Sep 27, 2017

WIP: running but not agreeing with other solver
jonathan-taylor
committed
c26a06b

Commits on Sep 28, 2017

Commits on Sep 29, 2017

Merge pull request #31 from jonathan-taylor/wide_matrix_QP
tibshirani
authored
726b917

Commits on Oct 4, 2017

rob fixed type=full option
tibshirani
committed
c0dab8c
rob type=full
tibshirani
committed
ae37fdc
rob fixed another type=full bug
tibshirani
committed
8f356ff
small change to current CRAN
tibshirani
committed
782290e

Commits on Oct 10, 2017

travis failing with install for shell reason -- maybe caps works better?
jonathan-taylor
committed
96e87ab
Merge remote-tracking branch 'upstream'
jonathan-taylor
committed
9d54c53
small cleanup of fixedLassoPoly
jonathan-taylor
committed
be257f1
rewriting fixedLassoPoly to allow for vector lambda (including unpenalized 0 coords)
jonathan-taylor
committed
25dce24
undefined reference to n
jonathan-taylor
committed
17eaff4
needed LaTeX escape chars
jonathan-taylor
committed
eb2a5e4
Merge pull request #32 from jonathan-taylor/master
jonathan-taylor
authored
334e007
added a full n>p example
jonathan-taylor
committed
43fbddf

Commits on Oct 11, 2017

Merge branch 'master' of github.com:selective-inference/R-software
jonathan-taylor
committed
151c21d
fix of signs for full
jonathan-taylor
committed
a6fcd6e
Merge pull request #33 from jonathan-taylor/master
tibshirani
authored
c5bf9f6
variable rename
jonathan-taylor
committed
8ad2f63
rob added unifTest.R
tibshirani
committed
749c136
rob modified unifTest.R
tibshirani
committed
9c7ee2b
rob modified unifTest.R
tibshirani
committed
85f7063

Commits on Oct 27, 2017

rob added unifTest.R

tibshirani
authored and
jonathan-taylor
committed
5f971f6
rob modified unifTest.R

tibshirani
authored and
jonathan-taylor
committed
42fc515
rob modified unifTest.R

tibshirani
authored and
jonathan-taylor
committed
781609d
NF: density functions for randomized LASSO
jonathan-taylor
committed
3da6c50
T to TRUE
jonathan-taylor
committed
e0eb1c6
updated wide C code -- can now do ridge as well as variable weights -- ready to solve randomized LASSO
jonathan-taylor
committed
b23b189
removing machine generated files
jonathan-taylor
committed
ffccea5
WIP: function to fit randomized lasso
jonathan-taylor
committed
2c0e3d6
check that solution is same as old code -- currently failing
jonathan-taylor
committed
2439d73
bug found in qp solver -- look at tests/test_QP.R
jonathan-taylor
committed
83b0426
allowing randomization to be 0 in solver
jonathan-taylor
committed
f13bce4
cosmetic edit
jonathan-taylor
committed
a6064f2
LASSO solver is right up to scale
jonathan-taylor
committed
2d685c0
Confidence intervals when estimate is negative
nicoballarini
authored
cc88b97
Confidence intervals when estimate is negative
nicoballarini
authored
7b23150
variable rename
jonathan-taylor
committed
928fee9
NF: density functions for randomized LASSO
jonathan-taylor
committed
2e9b11d
T to TRUE
jonathan-taylor
committed
c3b55f5
updated wide C code -- can now do ridge as well as variable weights -- ready to solve randomized LASSO
jonathan-taylor
committed
3748723
removing machine generated files
jonathan-taylor
committed
5a07245
WIP: function to fit randomized lasso
jonathan-taylor
committed
ede2816
check that solution is same as old code -- currently failing
jonathan-taylor
committed
32feb0e
bug found in qp solver -- look at tests/test_QP.R
jonathan-taylor
committed
0a99f5f
allowing randomization to be 0 in solver
jonathan-taylor
committed
242218e
cosmetic edit
jonathan-taylor
committed
360dcbd
LASSO solver is right up to scale
jonathan-taylor
committed
f5a3b96
BF: fixing Xsoln incorrectly set in wide solver
jonathan-taylor
committed
506a205
gradient not current?
jonathan-taylor
committed
0be39c4
print R2 from lm
jonathan-taylor
committed
3ee2a43
without linesearch we know agree with Adel's code at fixed mu
jonathan-taylor
committed
92812fc
fixing documentation
jonathan-taylor
committed
abf0824
checking comparison example again -- looks good
jonathan-taylor
committed
1658ecb
removing the offset_correction which was just for the comparison
jonathan-taylor
committed
e3049cf
Merge pull request #35 from nicoballarini/patch-1
jonathan-taylor
authored
32d4bd5
Merge pull request #34 from nicoballarini/patch-2
jonathan-taylor
authored
4e42c75

Commits on Oct 28, 2017

created the matrices for the affine transform, wrapper for calling Gaussian density
jonathan-taylor
committed
fc49dcd
unwanted files
jonathan-taylor
committed
d1dc82d
wrapper for conditional density
jonathan-taylor
committed
8ca2028
forming internal affine transform
jonathan-taylor
committed
07d7c6c
vectorized densities evaluate OK -- need to check results
jonathan-taylor
committed
9d1c79b
added laplace densities
jonathan-taylor
committed
5393dcd

Commits on Oct 31, 2017

using C-software submodule for selectiveInference/src
jonathan-taylor
committed
b637448
updating travis script
jonathan-taylor
committed
39b9675
variable rename
jonathan-taylor
committed
cd918b1
NF: density functions for randomized LASSO
jonathan-taylor
committed
ac2bbc8
T to TRUE
jonathan-taylor
committed
75c71dc
updated wide C code -- can now do ridge as well as variable weights -- ready to solve randomized LASSO
jonathan-taylor
committed
49d0b67
removing machine generated files
jonathan-taylor
committed
0dbc310
WIP: function to fit randomized lasso
jonathan-taylor
committed
7480310
check that solution is same as old code -- currently failing
jonathan-taylor
committed
38b2020
bug found in qp solver -- look at tests/test_QP.R
jonathan-taylor
committed
dc423de
allowing randomization to be 0 in solver
jonathan-taylor
committed
71ff6e3
cosmetic edit
jonathan-taylor
committed
de8874b
LASSO solver is right up to scale
jonathan-taylor
committed
07f738c
BF: fixing Xsoln incorrectly set in wide solver
jonathan-taylor
committed
2f1686e
gradient not current?
jonathan-taylor
committed
6444472
print R2 from lm
jonathan-taylor
committed
9f26844
without linesearch we know agree with Adel's code at fixed mu
jonathan-taylor
committed
9c7bfcc
fixing documentation
jonathan-taylor
committed
c2460a1
checking comparison example again -- looks good
jonathan-taylor
committed
ce15b68
removing the offset_correction which was just for the comparison
jonathan-taylor
committed
e01d8f1
created the matrices for the affine transform, wrapper for calling Gaussian density
jonathan-taylor
committed
eedf0fd
unwanted files
jonathan-taylor
committed
e8a94ef
wrapper for conditional density
jonathan-taylor
committed
b2de9d9
forming internal affine transform
jonathan-taylor
committed
6dfdd0d
vectorized densities evaluate OK -- need to check results
jonathan-taylor
committed
c175a8b
added laplace densities
jonathan-taylor
committed
c940537
BF: variable rename
jonathan-taylor
committed
206e236
BF: forgot to cd up directory
jonathan-taylor
committed
2917d6e
changing order of stop arguments
jonathan-taylor
committed
63e60df
changing order of stops
jonathan-taylor
committed
57d91d4
Merge pull request #36 from jonathan-taylor/master
tibshirani
authored
ff598d0
Merge branch 'master' into cython
jonathan-taylor
committed
46713f0
Merge pull request #37 from jonathan-taylor/master
jonathan-taylor
authored
232760d

Commits on Nov 1, 2017

Commits on Nov 2, 2017

WIP: have sample from optimization density, importance weight -- need to write pivot function
jonathan-taylor
committed
ae9aec8

Commits on Nov 3, 2017

Commits on Nov 5, 2017

pivots running but weights all zero
Jelena Markovic
committed
3ce4ea2
reorg
Jelena Markovic
committed
555f38c

Commits on Nov 6, 2017

segfault error
Jelena Markovic
committed
fdff629
bug in weights
Jelena Markovic
committed
68e27f4
condl bug fixed
Jelena Markovic
committed
0fa7c4e
added ci
Jelena Markovic
committed
8f99570
removed local import
jonathan-taylor
committed
9597497
Merge branch 'jelena-markovic-randomized_jelena'
jonathan-taylor
committed
a8d50a6

Commits on Nov 7, 2017

BF: segfault seems fixed -- ensured all functions have some order of ncase, nfeature
jonathan-taylor
committed
71b6586
jelena's inference functions
jonathan-taylor
committed
e0ee2a7
BF: seems ncase and nfeature were swapped in some places in C code
jonathan-taylor
committed
09db28a
minor cleanup -- doesn't seem to use conditional_density?
jonathan-taylor
committed
95b98c2
Merge pull request #1 from jonathan-taylor/master
Jelena Markovic
authored
a1ad181
coverages
Jelena Markovic
committed
07b6bb7
subgrad condition option added
Jelena Markovic
committed
1115635
bug
Jelena Markovic
committed
e2f4585
using some default arguments now -- bug in C code for densities
jonathan-taylor
committed
df777e3
dealing with zero length active set
Jelena Markovic
committed
ff34a52
Merge branch 'jelena-markovic-randomized_jelena'
jonathan-taylor
committed
1c538ed
now nut running after trying to set default arguments
jonathan-taylor
committed
ab91969
BF: estimate of sigma
jonathan-taylor
committed
5b4bc08
changing arguments
jonathan-taylor
committed
46856e1
BF: noise_scale was NULL
jonathan-taylor
committed
d1ebca3
seems to be running now with default arguments
jonathan-taylor
committed
0f1833f
forgot one bit of C code to use
jonathan-taylor
committed
8187b53
weights running in |E|+1 dimensions
Jelena Markovic
committed
6dae27c
BF: opt_transform linear_term was not cast as matrix in conditional_density
jonathan-taylor
committed
e90a8f1

Commits on Nov 8, 2017

working instance
Jelena Markovic
committed
2b49b94
export of randomized and man page
jonathan-taylor
committed
571cf82
adding resid to namespace
jonathan-taylor
committed
9adbe9b
fixing to make check pass
jonathan-taylor
committed
eda1482
param_stop
jonathan-taylor
committed
3ca4431
merged with smaller dimensional matrices
jonathan-taylor
committed
d838b07
For gaussian randomization, remove all p-dimensional matrices
jonathan-taylor
committed
7f8b8fd
fixing function signature
jonathan-taylor
committed
690f0ea
Merge pull request #39 from jonathan-taylor/master
tibshirani
authored
9e7a081
added mean and cov of a gaussian
Jelena Markovic
committed
6c1a123
transpose
Jelena Markovic
committed
d2a3b9b

Commits on Nov 9, 2017

working! bugs fixed. new sampler
Jelena Markovic
committed
e2efbdc
active set zero
Jelena Markovic
committed
e739255
removed set seed when gen randomization
Jelena Markovic
committed
b4fd8b5
amir sampler working with subgrad condition false as well
Jelena Markovic
committed
d2438c9
logistic
Jelena Markovic
committed
ce13171

Commits on Nov 10, 2017

E of zero length
Jelena Markovic
committed
157c1db
fixing man page so check passes
jonathan-taylor
committed
1dfbea8
Merge branch 'jelena-markovic-randomized_jelena'
jonathan-taylor
committed
fa08285
renaming file
jonathan-taylor
committed
b21260a
stop instead of print
jonathan-taylor
committed
183bc3a
Merge branch 'master' into logistic
jonathan-taylor
authored
4fc539d
WIP: adding logistic option, need to solve logistic problem
jonathan-taylor
committed
42cd878
Merge branch 'jelena-markovic-logistic'
jonathan-taylor
committed
69bf473
fixing empty inactive set indexing problem
jonathan-taylor
committed
af3b818
Merge pull request #41 from jonathan-taylor/master
tibshirani
authored
a2e32dc

Commits on Nov 15, 2017

rob chanegd NAMESPACE
tibshirani
committed
9de1b7c

Commits on Apr 9, 2018

Fixed handling of numeric variables in factorDesign
coreybrier
committed
a6c495b