Multinest #1034

ben18785 · 2020-02-12T12:30:30Z

Fixes #282.

…r multinest

codecov · 2020-02-18T12:18:43Z

Codecov Report

Merging #1034 (c3fad40) into master (9052e5d) will decrease coverage by 3.39%.
The diff coverage is 18.15%.

@@             Coverage Diff             @@
##            master    #1034      +/-   ##
===========================================
- Coverage   100.00%   96.60%   -3.40%     
===========================================
  Files           87       65      -22     
  Lines         8933     7047    -1886     
===========================================
- Hits          8933     6808    -2125     
- Misses           0      239     +239

Impacted Files	Coverage Δ
pints/_nested/_ellipsoid.py	`100.00% <ø> (ø)`
pints/_nested/_rejection.py	`100.00% <ø> (ø)`
pints/_nested/_multinest.py	`16.84% <16.84%> (ø)`
pints/_nested/__init__.py	`99.39% <71.42%> (-0.61%)`	⬇️
pints/io.py	`100.00% <0.00%> (ø)`
pints/_core.py	`100.00% <0.00%> (ø)`
pints/_util.py	`100.00% <0.00%> (ø)`
pints/noise.py	`100.00% <0.00%> (ø)`
pints/_logger.py	`100.00% <0.00%> (ø)`
pints/_log_pdfs.py	`100.00% <0.00%> (ø)`
... and 82 more

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 9052e5d...cdfba3f. Read the comment docs.

- created new getter - used getter in notebooks and test files

…nest

- added leafnodes getter

- also added checks for whether points lie in unit cube

- looks like way too many ellipsoids are being created though - also often get a "singular matrix" error (likely because ellipsoids are too small)

- also added f_s function - changed n_points > 50

ben18785 · 2021-03-09T23:26:31Z

Note to self:

include error handling for singular covariance matrix
clean up notebook and add ODE example(s)
look into dynamic enlargement factor as per ellipsoidal. Leave this out of this PR as it's an extension of the method really
clean up multinest function
add getter for underlying ellipsoid tree in multinest
correct algorithmic pseudocode to better reflect function (needs correcting again after new changes)
check enlargement factor hyperparameter use in old version
make ellipsoid update gap used
check number of hyperparameters: it said 6 but I can only find 3. I've settled on 4.

- added getter for ellipsoidtree - corrected pseudocode to be more representative of actual approach - added enlargement factor capacity - made ellipsoid gap used - corrected hyperparameters

Addresses #282

addresses #282

- also added max number of tries for while loop for kmeans addresses #282

- Error handling for singular matrix issue when forming bounding ellipse - Added back in recursion as it was used in the Matlab implementation we were following - Cleaned up the notebook and removed the logistic example as Goodwin is better - Improved docstrings for algo pseudocode Addresses #282

addresses #282

ben18785 added 12 commits February 10, 2020 12:19

added pseudocode for multinest and corrected pseudocode for ellipsoid

82348f1

corrected references in ellipsoid and rejection and added skeleton fo…

4a52bda

…r multinest

ask sketched out

7365948

alphabeticised multinest function ordering

dc5a450

skeleton of algorithm there

b3214d8

added notebook and skeleton of multinest complete (untested)

6d2f03c

created various docs files and tested doc building

f67bf4b

draft of multinest working with "unit-cubed" data

1293a5d

draft of multinest working on multimodal problem

70c45bf

Merge branch 'master' into i282-multinest

b88bc72

added transforming from/to unit cube

1f07e16

corrected logging error

c3fad40

MichaelClerx mentioned this pull request Feb 18, 2020

Sampler roadmap / overview #917

Closed

49 tasks

MichaelClerx added the new method label Nov 3, 2020

ben18785 added 15 commits March 2, 2021 15:38

Merge branch 'master' into i282-multinest

3d5815a

moved notebook

9ab1259

Merge branch 'master' into i282-multinest

3a6cc29

began to add tests

e2f0cc5

created sampler() method

fc8564e

- created new getter - used getter in notebooks and test files

Merge branch 'i1312-nested-controller-sampler-method' into i282-multi…

2437908

…nest

added ellipsoid class

80052b1

corrected logging issue for ellipsoidal nested sampling

2afc8f6

switched ellipsoidal nested sampling to use Ellipsoid class methods

b803fd7

added volume tests and corrected error with ellipsoidal nested sampling

1ca047b

added tests of ellipsoid sampling

13460da

added tests for min volume calc

40df3e7

added test for mahalanobis distance

394cc72

added some tests for multinest

c2b8149

started to think about ellipsoid sets

c90c2a1

ben18785 added 14 commits March 6, 2021 17:54

started to add h_k

8e684f1

first draft of ellipsoidtree

7c55a8d

overloaded addition to combine ellipsoids

61dfce4

- added leafnodes getter

removed ellipsoidset

e4100e7

cleaned up ellipsoidtree

7bf0ef2

cleaned up style issues

a602418

more style issues

64136f7

added safeguard against few points

8baa289

added tests of vs and vsk

c2b445a

- also added checks for whether points lie in unit cube

added ellipse plotting to multinest notebook

e0b71e9

commented out while loop for algorithmic oscillation in multinest

1af9adc

added sampling within ellipsoids method to ellipsoidtree

eefd07e

integrated ellipsoidtree in multinest

aaf3657

- looks like way too many ellipsoids are being created though - also often get a "singular matrix" error (likely because ellipsoids are too small)

typo in multinest

2a9e5a7

- also added f_s function - changed n_points > 50

ben18785 added 9 commits March 9, 2021 23:52

added max_recursion

e7cb8cd

#282 cleaned up multinest

2b45b98

- added getter for ellipsoidtree - corrected pseudocode to be more representative of actual approach - added enlargement factor capacity - made ellipsoid gap used - corrected hyperparameters

Cleans up docstrings for MultiNest

316c275

Addresses #282

multiNest notebook cleaned up

c084e01

addresses #282

removed recursion from MultiNest

6110010

- also added max number of tries for while loop for kmeans addresses #282

Added tests to Multinest

a9eed9f

addresses #282

Added test coverage for MultiNest

4f3c3a7

addresses #282

Cleaned up MultiNest notebook and ellipsoidal one

bb29984

addresses #282

ben18785 mentioned this pull request Mar 13, 2021

Created sampler() method within NestedController #1313

Closed

ben18785 added 2 commits March 13, 2021 13:28

Added @rcw5890 comment

532cd67

Merge branch 'master' into i282-multinest

93d9330

ben18785 marked this pull request as ready for review March 13, 2021 13:30

Corrected failing docs and copyright test for MultiNest

cdfba3f

addresses #282

ben18785 closed this Mar 13, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Multinest #1034

Multinest #1034

ben18785 commented Feb 12, 2020 •

edited by MichaelClerx

Loading

codecov bot commented Feb 18, 2020 •

edited

Loading

ben18785 commented Mar 9, 2021 •

edited

Loading

Multinest #1034

Multinest #1034

Conversation

ben18785 commented Feb 12, 2020 • edited by MichaelClerx Loading

codecov bot commented Feb 18, 2020 • edited Loading

Codecov Report

ben18785 commented Mar 9, 2021 • edited Loading

ben18785 commented Feb 12, 2020 •

edited by MichaelClerx

Loading

codecov bot commented Feb 18, 2020 •

edited

Loading

ben18785 commented Mar 9, 2021 •

edited

Loading