issues
search
stephenslab
/
mixsqp
R package for fast maximum-likelihood estimation of mixture proportions using SQP.
https://bit.ly/2NtYHWT
Other
11
stars
7
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
Compile error on Apple Silicon chip
#50
gaow
opened
1 year ago
3
solve(): system is singular
#49
william-denault
opened
2 years ago
8
Objective is -Inf in weighted EM
#48
william-denault
opened
2 years ago
1
irlba requires R >= 3.6.2
#47
mcolosimo-p4
closed
2 years ago
1
Refitting ash using mixsqp
#46
william-denault
closed
2 years ago
4
Solution is sensitive to random seed when low-rank approximation is used
#45
pcarbo
opened
4 years ago
1
Error when L matrix has one or more columns that are all zero
#44
pcarbo
closed
4 years ago
0
Add quadratic penalty to mixsqp objective
#43
pcarbo
closed
4 years ago
0
NPMLE example where mixsqp fails to converge
#42
pcarbo
closed
4 years ago
1
Add reduced Hessian output
#41
pcarbo
closed
4 years ago
1
Change behaviour of "tol.svd" for determining effective rank of L matrix
#40
pcarbo
opened
4 years ago
0
"Missing value" error
#39
willwerscheid
closed
4 years ago
2
Run one or more EM updates in mix-SQP "outer loop" for more reliable convergence
#38
pcarbo
closed
4 years ago
1
Not available to install from CRAN
#37
halasadi
closed
4 years ago
2
unused variable warning during installation
#36
willwerscheid
closed
4 years ago
2
Improve runtime when L has many columns
#35
pcarbo
closed
4 years ago
0
Add "log" argument to mixsqp interface
#34
halasadi
closed
5 years ago
5
Minor formatting proposal on strings 154 & 156
#33
kant
closed
5 years ago
1
Implement low-rank approximation to data matrix
#32
pcarbo
closed
4 years ago
2
Slow performance using 0.1-119
#31
jean997
opened
5 years ago
2
Example from smashr where mixsqp fails to converge
#30
pcarbo
closed
5 years ago
1
Add EM "pre-fitting" step to mixsqp method
#29
pcarbo
closed
5 years ago
1
Active set method does not terminate when all co-ordinates are positive
#28
pcarbo
closed
5 years ago
0
Adjust default for "maxiter.activeset" control parameter
#27
pcarbo
closed
5 years ago
0
Suggestion for documentation
#26
willwerscheid
closed
5 years ago
4
mixsqp can return (numerically) negative values
#25
willwerscheid
closed
5 years ago
5
Allow negative entries in L
#24
pcarbo
opened
5 years ago
0
Add step size to verbose=TRUE and "progress" return value in mixsqp
#23
pcarbo
closed
5 years ago
0
Remove check for step size being "too small"
#22
pcarbo
closed
6 years ago
0
Make backtracking line search more customizable
#21
pcarbo
closed
5 years ago
2
Handle likelihood matrices with a single column.
#20
willwerscheid
closed
6 years ago
2
Another (more difficult) example in which mixSQP fails to converge
#19
pcarbo
closed
6 years ago
1
Deal with columns of all zeros
#18
willwerscheid
closed
6 years ago
10
Example where mixsqp fails to converge for mashr
#17
pcarbo
closed
5 years ago
6
Idea: use a few iterations of EM to initialize mixsqp
#16
pcarbo
closed
6 years ago
0
Rename "data" to "progress" in mixsqp output
#15
pcarbo
closed
6 years ago
0
Add help("mixsqp-package")
#14
pcarbo
closed
6 years ago
0
Running on random uniform ensemble example
#13
youngseok-kim
closed
6 years ago
2
Consider normalizing eps separately for each row of L
#12
pcarbo
closed
6 years ago
0
Add "normalize.rows" option in mixsqp function
#11
pcarbo
closed
6 years ago
1
Add Jean's second examples as a test
#10
pcarbo
closed
6 years ago
1
Add checkUserInterrupt calls
#9
pcarbo
closed
6 years ago
0
Fix incorrect solution when initial estimate x0 yields an L*x0 with one or more very small entries
#8
pcarbo
closed
6 years ago
1
Move all mixsqp optimization parameters into a "control" argument
#7
pcarbo
closed
6 years ago
1
Add "Quick Start" to README
#6
pcarbo
closed
6 years ago
0
Allow for zero likelihoods
#5
pcarbo
closed
6 years ago
1
remove stray lines about ashr
#4
stephens999
closed
6 years ago
2
Make a few improvements to the vignette
#3
pcarbo
closed
6 years ago
0
Hide the old package repository
#2
pcarbo
closed
6 years ago
0
Feedback from Jason & Jean
#1
pcarbo
closed
6 years ago
0
Next