zhengruifeng commented on a change in pull request #25715:
[SPARK-28969][PYTHON][ML] OneVsRestParams parity between scala and python
URL: https://github.com/apache/spark/pull/25715#discussion_r323573836
##########
File path: python/pyspark/ml/classification.py
##########
@@ -1959,6 +1950,15 @@ def setParams(self, featuresCol="features",
labelCol="label", predictionCol="pre
kwargs = self._input_kwargs
return self._set(**kwargs)
+ @since("2.0.0")
+ def setClassifier(self, value):
+ """
+ Sets the value of :py:attr:`classifier`.
+
+ .. note:: Only LogisticRegression and NaiveBayes are supported now.
Review comment:
```python
In [1]: from pyspark.ml.linalg import Vectors
In [2]: from pyspark.sql import Row
In [3]: from pyspark.ml.classification import *
In [4]: svm = LinearSVC(maxIter=5, regParam=0.01)
In [5]: ovr = OneVsRest(classifier=svm)
In [6]: data_path = "data/mllib/sample_multiclass_classification_data.txt"
In [7]: df = spark.read.format("libsvm").load(data_path)
19/09/12 14:25:27 WARN LibSVMFileFormat: 'numFeatures' option not specified,
determining the number of features by going though the input. If you know the
number in advance, please specify it via 'numFeatures' option to avoid the
extra scan.
In [8]: model = ovr.fit(df)
19/09/12 14:25:38 WARN BLAS: Failed to load implementation from:
com.github.fommil.netlib.NativeSystemBLAS
19/09/12 14:25:38 WARN BLAS: Failed to load implementation from:
com.github.fommil.netlib.NativeRefBLAS
In [9]: model.models[0].coefficients
Out[9]: DenseVector([1.0653, 0.5648, 1.0451, 1.5259])
In [10]: model.getClassifier()
Out[10]: LinearSVC_1730373ceae7
```
one more thing, OVR now support more base models. So we can del this note.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
