Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Sep 3, 2020

• Ready for submission

  

  matan_ma

  committed
  b0fec5c

  Copy full SHA for b0fec5c

Commits on Sep 2, 2020

• generated submission. Before last (?) cleaning.

  

  matan_ma

  committed
  3cdd316

  Copy full SHA for 3cdd316

Commits on Sep 1, 2020

• Added the change. Reverting....

  

  matan_ma

  committed
  97b8bc7

  Copy full SHA for 97b8bc7

• Bad news. The test set has different ports than the train set, so the features generated by the one hot encoder are different than the ones generated by the test set.... I'll need to have a patch here

  

  matan_ma

  committed
  9f37e01

  Copy full SHA for 9f37e01

• XGBoost cv still running. Committing before removing the per-vessel unique ports

  

  matan_ma

  committed
  478ae6e

  Copy full SHA for 478ae6e

• Ok, the train-set dependent features won't work, since in the cross validation I also need to base these on each of the internal training sets. The good news are that the sparse matrix makes xgboos…

  Show description for 355c99b
  

  matan_ma

  committed
  355c99b

  Copy full SHA for 355c99b

• There's a bug with the features construction. Committing before fixing

  

  matan_ma

  committed
  9fd62e8

  Copy full SHA for 9fd62e8

Commits on Aug 31, 2020

• Had to restart. Most is ready

  

  matan_ma

  committed
  b5790c4

  Copy full SHA for b5790c4

• Ironing WinWard before submission

  

  matan_ma

  committed
  c23d5d1

  Copy full SHA for c23d5d1

Commits on Aug 30, 2020

• First results across several classifiers

  

  matan_ma

  committed
  770eedf

  Copy full SHA for 770eedf

• other features work o.k.

  

  matan_ma

  committed
  4af1541

  Copy full SHA for 4af1541

Commits on Aug 29, 2020

• Cleaning seems to be working...

  

  matan_ma

  committed
  6315c7e

  Copy full SHA for 6315c7e

• Mainly data cleaning

  

  matan_ma

  committed
  63a89d8

  Copy full SHA for 63a89d8

• Initial classification works

  

  matan_ma

  committed
  495e060

  Copy full SHA for 495e060

• More EDA

  

  matan_ma

  committed
  9293058

  Copy full SHA for 9293058

Commits on Aug 28, 2020

• More EDA

  

  matan_ma

  committed
  10d6ccc

  Copy full SHA for 10d6ccc

Commits on Aug 27, 2020

• Added wnwd homework

  

  matan_ma

  committed
  8aa8adf

  Copy full SHA for 8aa8adf

Commits on Aug 10, 2020

• File used for submission

  

  matan_ma

  committed
  5812dcf

  Copy full SHA for 5812dcf

• After tuning over the full training set. Before implementing oversampling only over the train set

  

  matan_ma

  committed
  941251c

  Copy full SHA for 941251c

Commits on Aug 9, 2020

• Partial train set - closed. Now running grid search over full train set

  

  matan_ma

  committed
  b281841

  Copy full SHA for b281841

Commits on Aug 8, 2020

• Before trying to integrate smote into the GridSearchCV

  

  matan_ma

  committed
  ff0b6e0

  Copy full SHA for ff0b6e0

• Smote improved

  

  matan_ma

  committed
  d2ceb79

  Copy full SHA for d2ceb79

• Extracted mis-classified across all raw training set and visualized.

  

  matan_ma

  committed
  0677836

  Copy full SHA for 0677836

Commits on Aug 7, 2020

• Committing before adding two features: intersection of y and z, and mean distance between them

  

  matan_ma

  committed
  77f65b0

  Copy full SHA for 77f65b0

• Implemented grid_search over unique patients. Some plots to realize class 1 misses

  

  matan_ma

  committed
  a8ddcfe

  Copy full SHA for a8ddcfe

Commits on Aug 6, 2020

• Removed duplicates, added class weights in xgboost in /ClewHomeAssignment_more_experiments.ipynb

  

  matan_ma

  committed
  61b7f0b

  Copy full SHA for 61b7f0b

• Added ClewMedical/ClewHomeAssignment_more_experiments.ipynb

  

  matan_ma

  committed
  c2460e2

  Copy full SHA for c2460e2

Commits on Jul 10, 2020

• Content based filtering seems to work. Committing before making some amendments

  

  matan_ma

  committed
  c6a77d0

  Copy full SHA for c6a77d0

Commits on Jul 8, 2020

• initial simulation runs

  

  matan_ma

  committed
  3c3f962

  Copy full SHA for 3c3f962

• working on Ebay's assignment code

  

  matan_ma

  committed
  dad1faf

  Copy full SHA for dad1faf

Commits on Jul 7, 2020

• Submitted version

  

  matan_ma

  committed
  b9f6483

  Copy full SHA for b9f6483

• Committing after cleaning the notebook, before removing the training over entire train set.

  

  matan_ma

  committed
  a063977

  Copy full SHA for a063977

Commits on Jul 6, 2020

• Committing before deleting all trailing cells and submitting

  

  matan_ma

  committed
  33552fa

  Copy full SHA for 33552fa

• Merge branch 'master' of https://github.com/matanma/ML

  

  matan_ma

  committed
  9d6a776

  Copy full SHA for 9d6a776

• Added redundant notebooks

  

  matan_ma

  committed
  0eed4fd

  Copy full SHA for 0eed4fd

Pagination

Previous

Next