search

VarIr / scikit-hubness

A Python package for hubness analysis and high-dimensional data mining
BSD 3-Clause "New" or "Revised" License
44 stars 10 forks source link

FIX hubness estimation error on negative nn indices #59

Closed VarIr closed 4 years ago

VarIr commented 4 years ago

Hubness estimation can error on negative neighbor indices that can occur when ANN does not find enough neighbors. These are now filtered away for hubness estimation. Also slightly change how to select the memory available to puffinn.

codecov-io commented 4 years ago

Codecov Report

Merging #59 into master will decrease coverage by 0.04%. The diff coverage is 88.23%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master      #59      +/-   ##
==========================================
- Coverage   99.17%   99.13%   -0.05%     
==========================================
  Files          55       55              
  Lines        4491     4505      +14     
  Branches      494      496       +2     
==========================================
+ Hits         4454     4466      +12     
- Misses         18       19       +1     
- Partials       19       20       +1
Impacted Files Coverage Δ
skhubness/analysis/tests/test_estimation.py 99.47% <100%> (+0.02%) :arrow_up:
skhubness/analysis/estimation.py 99.57% <100%> (ø) :arrow_up:
skhubness/neighbors/lsh.py 96.41% <33.33%> (-1.03%) :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 7650474...cfa0c8e. Read the comment docs.