search

andrewhooker / PopED

Population Experimental Design (PopED) in R
https://andrewhooker.github.io/PopED/
GNU Lesser General Public License v3.0
33 stars 21 forks source link

Try Cholesky Decomposition before falling back to SVD #19

Closed martin-gmx closed 6 years ago

martin-gmx commented 6 years ago

SVD is sooo slow, so it should be okay to at least "try" the Cholesky Decomposition before. It seems that the Cholesky Decomposition is also not that much longer than calculating the Determinant, which would be an alternative method to test if the matrix is invertible.

codecov[bot] commented 6 years ago

Codecov Report

Merging #19 into master will increase coverage by 0.01%. The diff coverage is 100%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master      #19      +/-   ##
==========================================
+ Coverage    53.1%   53.11%   +0.01%     
==========================================
  Files         144      144              
  Lines        9654     9652       -2     
==========================================
  Hits         5127     5127              
+ Misses       4527     4525       -2
Impacted Files Coverage Δ
R/inv.R 100% <100%> (ø) :arrow_up:
R/ed_laplace_ofv.R 26.64% <0%> (+0.59%) :arrow_up:

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 5c6e71a...9441afd. Read the comment docs.