rhiever/tpot: Sparse matrix support, early stopping, and checkpointing
Creators
- 1. University of Pennsylvania
- 2. @EpistasisLab
- 3. Knowledge Lab
- 4. Michigan State University
- 5. Telefonica I+D
- 6. Transfix
- 7. @betteroutcomes
Description
TPOT now supports sparse matrices with a new built-in TPOT configurations, "TPOT sparse". We are using a custom OneHotEncoder implementation that supports missing values and continuous features.
We have added an "early stopping" option for stopping the optimization process if no improvement is made within a set number of generations. Look up the
early_stop
parameter to access this functionality.TPOT now reduces the number of duplicated pipelines between generations, which saves you time during the optimization process.
TPOT now supports custom scoring functions via the command-line mode.
We have added a new optional argument,
periodic_checkpoint_folder
, that allows TPOT to periodically save the best pipeline so far to a local folder during optimization process.TPOT no longer uses
sklearn.externals.joblib
whenn_jobs=1
to avoid the potential freezing issue that scikit-learn suffers from.We have added
pandas
as a dependency to read input datasets instead ofnumpy.recfromcsv
. NumPy'srecfromcsv
function is unable to parse datasets with complex data types.Fixed a bug that
DEFAULT
in the parameter(s) of nested estimator raisesKeyError
when exporting pipelines.Fixed a bug related to setting
random_state
in nested estimators. The issue would happen with pipeline withSelectFromModel
(ExtraTreesClassifier
as nested estimator) orStackingEstimator
if nested estimator hasrandom_state
parameter.Fixed a bug in the missing value imputation function in TPOT to impute along columns instead rows.
Refined input checking for sparse matrices in TPOT.
Files
rhiever/tpot-v0.9.zip
Files
(2.5 MB)
Name | Size | Download all |
---|---|---|
md5:e76727b785d262f126adc346a29a5a5f
|
2.5 MB | Preview Download |
Additional details
Related works
- Is supplement to
- https://github.com/rhiever/tpot/tree/v0.9 (URL)