Slope#2

Open

jonathan-taylor wants to merge 673 commits into

snigdhagit:SLOPEfrom

jonathan-taylor:SLOPE

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

Commits on Nov 17, 2017

Merge branch 'adjusted_MLE' of https://github.com/snigdhagit/selective-inference into adjusted_MLE

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
commit test for non-orthogonal LASSO

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
added test for comparison of variances

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Nov 18, 2017

commit changes

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Nov 19, 2017

new computation for observed Fisher info

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
checked non orthogonal lasso p 200

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Nov 20, 2017

added test to compute univariate MLE based on E maps

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Nov 21, 2017

test BH with orthogonal design

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
added test for BH screening

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
commit changes

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
added additional constraint in BH

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Nov 24, 2017

commit comparsion with bootstrapped variance

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
changed map for sigma not equal to 1.

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Nov 25, 2017

added observed fisher info as a function of target

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
return inv of hessian

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
changed previous tests for MLE

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
commit changes

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Nov 26, 2017

commit changes

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Nov 28, 2017

check sd from bootstrapped pivot

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Nov 30, 2017

tried estimating var based on a sample from selective distribution

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Dec 4, 2017

estimating sigma using glmnet

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
added arguments for relative risk computation

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
changes

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Dec 5, 2017

merging with snigdha recent work
jonathan-taylor
committed
print correct bias
jonathan-taylor
committed
est sigma, lambda from CV.glmnet and option for independent estimator

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
found bug in offset term while setting implied mean

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Dec 6, 2017

added risk function

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
compute rel risks

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Dec 7, 2017

changed scale of glmnet

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
cleaned risk

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
added comparison of estimators appended with zeros

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
commit additions of nonrand LASSO

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Dec 8, 2017

added risk comparisons

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
added coverage

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
commit all changes

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed
commit changes

Snigdha Panigrahi
authored and
Snigdha Panigrahi
committed

Commits on Dec 9, 2017

commit change
Snigdha Panigrahi
committed

Commits on Dec 10, 2017

called sim.xy
Snigdha Panigrahi
committed
commit merge
Snigdha Panigrahi
committed
getting coefs for all possible combinations of lambda and gamma
Snigdha Panigrahi
committed
tuned version of relaxed lasso
Snigdha Panigrahi
committed
fix sigma and change coef in dgp
Snigdha Panigrahi
committed
return tuned relaxed lasso est
Snigdha Panigrahi
committed
commit all changes
Snigdha Panigrahi
committed

Commits on Dec 11, 2017

corrected glmnet
Snigdha Panigrahi
committed
added both lambda min and 1se in glmnet
Snigdha Panigrahi
committed
comparisons for tuned estimator against sel MLE
Snigdha Panigrahi
committed
extract the tuned lambda in best subset and use it for rand LASSO
Snigdha Panigrahi
committed
tried tuning randomized LASSO
Snigdha Panigrahi
committed

Commits on Dec 12, 2017

use external validation set to tune lambda
Snigdha Panigrahi
committed
added tuned LASSO est
Snigdha Panigrahi
committed
added screening power
Snigdha Panigrahi
committed
added coverages
Snigdha Panigrahi
committed
added false positives to screening attributes
Snigdha Panigrahi
committed
added inferential power
Snigdha Panigrahi
committed

Commits on Dec 13, 2017

corrected power
Snigdha Panigrahi
committed
commit all changes
Snigdha Panigrahi
committed
a small correction
Snigdha Panigrahi
committed

Commits on Dec 14, 2017

merging
jonathan-taylor
committed

Commits on Dec 15, 2017

centered y and added partial risks
Snigdha Panigrahi
committed
fixed scales in partial risk
Snigdha Panigrahi
committed

Commits on Dec 16, 2017

Commits on Dec 17, 2017

updated changes of target and tuning
Snigdha Panigrahi
committed
added debiased target
Snigdha Panigrahi
committed
commit changes so far
Snigdha Panigrahi
committed
reverted back to fixing sigma as per snr in sim.xy
Snigdha Panigrahi
committed
normalized y's by sigma_est
Snigdha Panigrahi
committed
changed scale of relaxed lasso tuned
Snigdha Panigrahi
committed

Commits on Dec 18, 2017

added target in selection map
Snigdha Panigrahi
committed
bootstrap intervals very long for high dims
Snigdha Panigrahi
committed
commit changes
Snigdha Panigrahi
committed
installed best subset package-- enlist is changed
Snigdha Panigrahi
committed

Commits on Dec 19, 2017

commit before switch
Snigdha Panigrahi
committed
commit before switch
Snigdha Panigrahi
committed

Commits on Dec 20, 2017

run qq-norm diagnostic
Snigdha Panigrahi
committed
commit before switch
Snigdha Panigrahi
committed

Commits on Dec 21, 2017

Commits on Dec 28, 2017

coverage wo bootstrap
Snigdha Panigrahi
committed
commit changes
Snigdha Panigrahi
committed

Commits on Dec 29, 2017

selective coverage improving in high dimensions
Snigdha Panigrahi
committed
clean-up of code
Snigdha Panigrahi
committed
removed bootstrapping for interval construction
Snigdha Panigrahi
committed
more clean up and tests
Snigdha Panigrahi
committed
commit before switch
Snigdha Panigrahi
committed

Commits on Feb 7, 2018

Commits on Feb 8, 2018

Commits on Feb 9, 2018

Commits on Feb 13, 2018

Commits on Feb 14, 2018

WIP: writing highdim gaussian randomization version
jonathan-taylor
committed

Commits on Feb 16, 2018

comparing highdim to more general with decomposition
jonathan-taylor
committed

Commits on Feb 17, 2018

adding logdens_transform
jonathan-taylor
committed

Commits on Feb 22, 2018

edits to take into account given conditional_cov and mean
jonathan-taylor
committed

Commits on Feb 27, 2018

Commits on Mar 2, 2018

sign change in target_lin
Snigdha Panigrahi
committed
match with selective UMVU computation
Snigdha Panigrahi
committed

Commits on Mar 7, 2018

commit changes in test
Snigdha Panigrahi
committed
fix conflicts
Snigdha Panigrahi
committed

Commits on Mar 13, 2018

Commits on Mar 14, 2018

Commits on Mar 15, 2018

BF: debiasing_matrix was returning none
jonathan-taylor
committed

Commits on Mar 17, 2018

commit changes
Snigdha Panigrahi
committed
resolve conflict
Snigdha Panigrahi
committed

Commits on Mar 19, 2018

clean tests
Snigdha Panigrahi
committed
adding new test
Snigdha Panigrahi
committed

Commits on Mar 20, 2018

updated test
Snigdha Panigrahi
committed
added a return of indep est to the func. selective_MLE
Snigdha Panigrahi
committed
finished adding risks
Snigdha Panigrahi
committed

Commits on Mar 22, 2018

tuned lasso and lasso in python not having same solns
Snigdha Panigrahi
committed
comparison of risks
Snigdha Panigrahi
committed
removed redundant print
Snigdha Panigrahi
committed
commit changes in test
Snigdha Panigrahi
committed

Commits on Mar 27, 2018

reorganized test
Snigdha Panigrahi
committed
commit changes
Snigdha Panigrahi
committed
added inferential powers
Snigdha Panigrahi
committed
correction in power
Snigdha Panigrahi
committed
adding Liu et al to library
jonathan-taylor
committed

Commits on Mar 31, 2018

coverage for debiased target is short of target
Snigdha Panigrahi
committed

Commits on Apr 1, 2018

coverage falling short of target for debiased
Snigdha Panigrahi
committed

Commits on Apr 2, 2018

double requirement
jonathan-taylor
committed

Commits on Apr 3, 2018

Commits on Apr 7, 2018

Commits on Apr 8, 2018

restructured test and rectified scaling of target
Snigdha Panigrahi
committed

Commits on Apr 10, 2018

create plots + save results
Snigdha Panigrahi
committed
tried to tune randomization scale
Snigdha Panigrahi
committed
tolerance arguments for selective mle
jonathan-taylor
committed

Commits on Apr 11, 2018

scale of LASSO in python made to agree with glmnet
Snigdha Panigrahi
committed
corrected scales and reduced length of grid search for tuning randomized LASSO for selected targets
Snigdha Panigrahi
committed
corrected scales in selected target
Snigdha Panigrahi
committed
randomized LASSO est seems comparable in risk to tuned estimators when p>n
Snigdha Panigrahi
committed
added function fixedLassoInf from R
Snigdha Panigrahi
committed
corrected arguments that are passed to fixedLassoInf
Snigdha Panigrahi
committed
calling R package for Lee inference, also for selected targets
Snigdha Panigrahi
committed
add power post BH filter
Snigdha Panigrahi
committed

Commits on Apr 12, 2018

added fdr and power post passing p-values through BH sieve
Snigdha Panigrahi
committed
removed some unnecessary prints
Snigdha Panigrahi
committed
some more reorganization
Snigdha Panigrahi
committed
making notion of power consistent across tests
Snigdha Panigrahi
committed
organizing output
Snigdha Panigrahi
committed
generate outputs
Snigdha Panigrahi
committed
adding average selected size and discoveries
Snigdha Panigrahi
committed
run test for a grid of snr values
Snigdha Panigrahi
committed

Commits on Apr 13, 2018

add output files
Snigdha Panigrahi
committed
added SLOPE to current working branch
Snigdha Panigrahi
committed
test for SLOPE
Snigdha Panigrahi
committed
cleaned up test for SLOPE
Snigdha Panigrahi
committed

Commits on Apr 15, 2018

check soln of randomized SLOPE
Snigdha Panigrahi
committed
detect clusters and rearange subgradient in decreasing order
Snigdha Panigrahi
committed

Commits on Apr 16, 2018

created a temporary class for SLOPE for now
Snigdha Panigrahi
committed
removed bug in fixing feasible point
Snigdha Panigrahi
committed
removed a sign bug from score_linear
Snigdha Panigrahi
committed
removed print statements, coverage is short of target
Snigdha Panigrahi
committed
removed sign bug from opt_linear_term
Snigdha Panigrahi
committed
still coverage is short of target
Snigdha Panigrahi
committed

Commits on Apr 17, 2018

add output files
Snigdha Panigrahi
committed
commit results where tuning parameter is tuned using selective mle
Snigdha Panigrahi
committed

Commits on Apr 20, 2018

adding log posterior parametrized using sel_MLE in query
Snigdha Panigrahi
committed
added a missing term
Snigdha Panigrahi
committed
changed name of func
Snigdha Panigrahi
committed

Commits on Apr 24, 2018

logistic debiased liu tests
jonathan-taylor
committed

Commits on Apr 26, 2018

need to figure out what score is
Snigdha Panigrahi
committed

Commits on Apr 27, 2018

commit changes so far
Snigdha Panigrahi
committed
BH needs debugging
Snigdha Panigrahi
committed
affine version of barrier problem
jonathan-taylor
committed
updated of C software for barrier affine
jonathan-taylor
committed

Commits on Apr 28, 2018

commit changes
Snigdha Panigrahi
committed
BH is not working yet
Snigdha Panigrahi
committed
commit changes in branch before switch
Snigdha Panigrahi
committed

Commits on May 1, 2018

running slope with affine constraints, barrier code written in python
Snigdha Panigrahi
committed
coverage for SLOPE looks good
Snigdha Panigrahi
committed

Commits on May 2, 2018