Error: AttributeError: 'DataFrame' object has no attribute '_jdf'

Question

I want to perform k-fold cross validation using pyspark to finetune the parameters and I\'m using pyspark.ml. I am getting Attribute Error.

AttributeError: \'DataFrame\'

Answer 1

If a metric evaluation error you probably:

1. Transformed using Spark on test set properly, then peeked using Pandas DF.

# Spark model, transformed test, converted to pandas df
predictions = model.transform(test)
predDF = predictions.toPandas()
predDF.head()

2. Then tried:

eval_acc = MulticlassClassificationEvaluator(
            labelCol='Label_index',
            predictionCol='prediction',
            metricName='accuracy'
)

# Evaluate Performance
acc = eval_acc.evaluate(predDF) # Error
print(f"accuracy: {acc}")

I forgot predDF is a Pandas DataFrame. Needed predictions because its a Spark Dataframe.

acc = eval_acc.evaluate(predictions) # Works
print(f"accuracy: {acc}")