Closed iamDecode closed 4 years ago
Merging #15 into master will not change coverage. The diff coverage is
100%
.
@@ Coverage Diff @@
## master #15 +/- ##
=====================================
Coverage 100% 100%
=====================================
Files 9 11 +2
Lines 376 411 +35
=====================================
+ Hits 376 411 +35
Impacted Files | Coverage Δ | |
---|---|---|
sklearn_pmml_model/linear_model/__init__.py | 100% <ø> (ø) |
:arrow_up: |
sklearn_pmml_model/naive_bayes/implementations.py | 100% <100%> (ø) |
|
sklearn_pmml_model/naive_bayes/__init__.py | 100% <100%> (ø) |
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 309a659...103322c. Read the comment docs.
This PR adds basic support for Naive Bayes. Currently only the Gaussian distribution is supported. Categorical support is implemented by one hot encoding features, just like done for linear models. Results are consistent with predictions in R with
naiveBayes
from thee1071
library.Support for the
PoissonDistribution
is as far as I know not possible in scikit-learn. The different scikit-learn Naive Bayes implementations (Multinomial, Complement, Bernoulli, Categorical) still need to be implemented, but they are less relevant as they only apply to categorical (only) data sets. It will need some investigation to check how these models would be represented in PMML.