You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Raghuvarran V H (Jira)" <ji...@apache.org> on 2020/06/22 11:23:00 UTC
[jira] [Created] (SPARK-32050) GBTClassifier not working with
OnevsRest
Raghuvarran V H created SPARK-32050:
---------------------------------------
Summary: GBTClassifier not working with OnevsRest
Key: SPARK-32050
URL: https://issues.apache.org/jira/browse/SPARK-32050
Project: Spark
Issue Type: Bug
Components: ML
Affects Versions: 2.4.0
Environment: spark 2.4.0
Reporter: Raghuvarran V H
I am trying to use GBT classifier for multi class classification using OnevsRest
from pyspark.ml.classification import MultilayerPerceptronClassifier,OneVsRest,GBTClassifier
from pyspark.ml import Pipeline,PipelineModel
lr = GBTClassifier(featuresCol='features', labelCol='label', predictionCol='prediction', maxDepth=5, maxBins=32, minInstancesPerNode=1, \
minInfoGain=0.0, maxMemoryInMB=256, cacheNodeIds=False, checkpointInterval=10, lossType='logistic', maxIter=20,\
stepSize=0.1, seed=None, subsamplingRate=1.0, featureSubsetStrategy='auto')
classifier = OneVsRest(featuresCol='features', labelCol='label', predictionCol='prediction', \
classifier=lr, weightCol=None, parallelism=1)
pipeline = Pipeline(stages=[str_indxr,ohe,vecAssembler,normalizer,classifier])
model = pipeline.fit(train_data)
When I try this I get this error:
/opt/cloudera/parcels/CDH-6.3.1-1.cdh6.3.1.p0.1470567/lib/spark/python/pyspark/ml/classification.py in _fit(self, dataset)
1800 classifier = self.getClassifier()
1801 assert isinstance(classifier, HasRawPredictionCol),\
-> 1802 "Classifier %s doesn't extend from HasRawPredictionCol." % type(classifier)
1803
1804 numClasses = int(dataset.agg(\{labelCol: "max"}).head()["max("+labelCol+")"]) + 1
AssertionError: Classifier <class 'pyspark.ml.classification.GBTClassifier'> doesn't extend from HasRawPredictionCol.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org