Model Selection - Cross Validation


In [ ]:
# import the packages
from pyspark.ml import Pipeline
from pyspark.ml.classification import LogisticRegression
from pyspark.ml.evaluation import BinaryClassificationEvaluator
from pyspark.ml.feature import HashingTF, Tokenizer
from pyspark.ml.tuning import CrossValidator, ParamGridBuilder

In [ ]:
# Prepare training documents, which are labeled.
training = spark.createDataFrame([
    (0, "a b c d e spark", 1.0),
    (1, "b d", 0.0),
    (2, "spark f g h", 1.0),
    (3, "hadoop mapreduce", 0.0),
    (4, "b spark who", 1.0),
    (5, "g d a y", 0.0),
    (6, "spark fly", 1.0),
    (7, "was mapreduce", 0.0),
    (8, "e spark program", 1.0),
    (9, "a e c l", 0.0),
    (10, "spark compile", 1.0),
    (11, "hadoop software", 0.0)
], ["id", "text", "label"])

In [ ]:
# Configure an ML pipeline, which consists of tree stages: tokenizer, hashingTF, and lr.
tokenizer = Tokenizer(inputCol="text", outputCol="words")
hashingTF = HashingTF(inputCol=tokenizer.getOutputCol(), outputCol="features")
lr = LogisticRegression(maxIter=10)
pipeline = Pipeline(stages=[tokenizer, hashingTF, lr])

In [ ]:
# We now treat the Pipeline as an Estimator, wrapping it in a CrossValidator instance.
# This will allow us to jointly choose parameters for all Pipeline stages.
# A CrossValidator requires an Estimator, a set of Estimator ParamMaps, and an Evaluator.
# We use a ParamGridBuilder to construct a grid of parameters to search over.
# With 3 values for hashingTF.numFeatures and 2 values for lr.regParam,
# this grid will have 3 x 2 = 6 parameter settings for CrossValidator to choose from.
paramGrid = ParamGridBuilder() \
    .addGrid(hashingTF.numFeatures, [10, 100, 1000]) \
    .addGrid(lr.regParam, [0.1, 0.01]) \
    .build()

In [ ]:
crossval = CrossValidator(estimator=pipeline,
                          estimatorParamMaps=paramGrid,
                          evaluator=BinaryClassificationEvaluator(),
                          numFolds=2)  # use 3+ folds in practice

In [ ]:
# Run cross-validation, and choose the best set of parameters.
cvModel = crossval.fit(training)

In [ ]:
# Prepare test documents, which are unlabeled.
test = spark.createDataFrame([
    (4, "spark i j k"),
    (5, "l m n"),
    (6, "mapreduce spark"),
    (7, "apache hadoop")
], ["id", "text"])

In [ ]:
# Make predictions on test documents. cvModel uses the best model found (lrModel).
prediction = cvModel.transform(test)
selected = prediction.select("id", "text", "probability", "prediction")
for row in selected.collect():
    print(row)