spectra-cluster / spectra-cluster-2

A complete re-write of the spectra-cluster codebase
1 stars 2 forks source link

Optimised Kendall Tau correlation #72

Closed jgriss closed 5 years ago

jgriss commented 5 years ago

Main changes are:

Other optimisations I tested, including not computing ties anymore actually didn't improve the performance at all. Also here, the improvements are very small. Nevertheless, since it does optimize the code a bit I'd suggest to leave them.

codecov-io commented 5 years ago

Codecov Report

Merging #72 into develop will decrease coverage by <.01%. The diff coverage is 92.3%.

Impacted file tree graph

@@             Coverage Diff             @@
##             develop    #72      +/-   ##
===========================================
- Coverage      61.01%    61%   -0.01%     
- Complexity       396    398       +2     
===========================================
  Files             59     60       +1     
  Lines           2006   2003       -3     
  Branches         259    259              
===========================================
- Hits            1224   1222       -2     
  Misses           672    672              
+ Partials         110    109       -1
Impacted Files Coverage Δ Complexity Δ
.../org/spectra/cluster/tools/SpectraClusterTool.java 0% <ø> (ø) 0 <0> (ø) :arrow_down:
...lter/binaryspectrum/FractionTicFilterFunction.java 96.29% <100%> (+0.29%) 7 <0> (+1) :arrow_up:
...n/java/org/spectra/cluster/similarity/IntPair.java 66.66% <66.66%> (ø) 2 <2> (?)
...luster/similarity/CombinedFisherIntensityTest.java 71.92% <91.66%> (-1.41%) 9 <4> (ø)
...pectra/cluster/similarity/KendallsCorrelation.java 96.92% <95.23%> (+2.63%) 14 <14> (-1) :arrow_down:

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 5e2c5f5...398d656. Read the comment docs.