In [1]:

    

from pyspark.sql import SparkSession

spark = SparkSession \
    .builder \
    .appName("Python Spark Random Forest Regression") \
    .config("spark.some.config.option", "some-value") \
    .getOrCreate()


    

In [3]:

    

df = spark.read.format('com.databricks.spark.csv').\
                               options(header='true', \
                               inferschema='true').load("../data/WineData.csv",header=True);


    

In [4]:

    

from pyspark.ml.linalg import Vectors # !!!!caution: not from pyspark.mllib.linalg import Vectors
from pyspark.ml import Pipeline
from pyspark.ml.feature import IndexToString,StringIndexer, VectorIndexer
from pyspark.ml.classification import DecisionTreeClassifier
from pyspark.ml.tuning import CrossValidator, ParamGridBuilder
from pyspark.ml.evaluation import MulticlassClassificationEvaluator


    

In [5]:

    

def transData(data):
    return data.rdd.map(lambda r: [Vectors.dense(r[:-1]),r[-1]]).toDF(['features','label'])


    

In [6]:

    

data = transData(df)


    

    




/Users/mingchen/anaconda2/lib/python2.7/site-packages/pytz/__init__.py:29: UserWarning: Module argparse was already imported from /Users/mingchen/anaconda2/lib/python2.7/argparse.pyc, but /Users/mingchen/anaconda2/lib/python2.7/site-packages/argparse-1.4.0-py2.7.egg is being added to sys.path
  from pkg_resources import resource_stream

In [7]:

    

data.show(3)


    

    




+--------------------+-----+
|            features|label|
+--------------------+-----+
|[7.4,0.7,0.0,1.9,...|    5|
|[7.8,0.88,0.0,2.6...|    5|
|[7.8,0.76,0.04,2....|    5|
+--------------------+-----+
only showing top 3 rows

In [8]:

    

# Index labels, adding metadata to the label column
labelIndexer = StringIndexer(inputCol='label',
                             outputCol='indexedLabel').fit(data)
labelIndexer.transform(data).show(6)


    

    




+--------------------+-----+------------+
|            features|label|indexedLabel|
+--------------------+-----+------------+
|[7.4,0.7,0.0,1.9,...|    5|         0.0|
|[7.8,0.88,0.0,2.6...|    5|         0.0|
|[7.8,0.76,0.04,2....|    5|         0.0|
|[11.2,0.28,0.56,1...|    6|         1.0|
|[7.4,0.7,0.0,1.9,...|    5|         0.0|
|[7.4,0.66,0.0,1.8...|    5|         0.0|
+--------------------+-----+------------+
only showing top 6 rows

In [9]:

    

# Automatically identify categorical features, and index them.
# Set maxCategories so features with > 4 distinct values are treated as continuous.
featureIndexer =VectorIndexer(inputCol="features", \
                              outputCol="indexedFeatures", \
                              maxCategories=4).fit(data)

featureIndexer.transform(data).show(6)


    

    




+--------------------+-----+--------------------+
|            features|label|     indexedFeatures|
+--------------------+-----+--------------------+
|[7.4,0.7,0.0,1.9,...|    5|[7.4,0.7,0.0,1.9,...|
|[7.8,0.88,0.0,2.6...|    5|[7.8,0.88,0.0,2.6...|
|[7.8,0.76,0.04,2....|    5|[7.8,0.76,0.04,2....|
|[11.2,0.28,0.56,1...|    6|[11.2,0.28,0.56,1...|
|[7.4,0.7,0.0,1.9,...|    5|[7.4,0.7,0.0,1.9,...|
|[7.4,0.66,0.0,1.8...|    5|[7.4,0.66,0.0,1.8...|
+--------------------+-----+--------------------+
only showing top 6 rows

In [10]:

    

# Train a DecisionTree model
dTree = DecisionTreeClassifier(labelCol='indexedLabel', featuresCol='indexedFeatures')


    

In [48]:

    

# Convert indexed labels back to original labels.
labelConverter = IndexToString(inputCol="prediction", outputCol="predictedLabel",
                               labels=labelIndexer.labels)


    

In [49]:

    

# Chain indexers and tree in a Pipeline
pipeline = Pipeline(stages=[labelIndexer, featureIndexer, dTree,labelConverter])


    

In [54]:

    

# Split the data into training and test sets (30% held out for testing)
(trainingData, testData) = data.randomSplit([0.7, 0.3])


    

In [56]:

    

# Train model.  This also runs the indexers.
model = pipeline.fit(trainingData)


    

In [58]:

    

# Make predictions.
predictions = model.transform(testData)


    

In [59]:

    

# Select example rows to display.
predictions.select("features","label","predictedLabel").show(5)


    

    




+--------------------+-----+--------------+
|            features|label|predictedLabel|
+--------------------+-----+--------------+
|[5.0,0.38,0.01,1....|    6|             6|
|[5.0,0.4,0.5,4.3,...|    6|             6|
|[5.0,0.74,0.0,1.2...|    6|             6|
|[5.0,1.04,0.24,1....|    5|             7|
|[5.1,0.42,0.0,1.8...|    7|             6|
+--------------------+-----+--------------+
only showing top 5 rows

In [66]:

    

# Search through decision tree's maxDepth parameter for best model
paramGrid = ParamGridBuilder().addGrid(dTree.maxDepth, [2,3,4,5,6,7]).build()


    

In [67]:

    

# Set F-1 score as evaluation metric for best model selection
evaluator = MulticlassClassificationEvaluator(labelCol='indexedLabel',
                                              predictionCol='prediction', metricName='f1')


    

In [68]:

    

# Set up 3-fold cross validation
crossval = CrossValidator(estimator=pipeline,
                          estimatorParamMaps=paramGrid,
                          evaluator=evaluator,
                          numFolds=3)


    

In [69]:

    

# Train model.  This also runs the indexers.
dTree_model = crossval.fit(data)


    

In [74]:

    

# Fetch best model
tree_model = dTree_model.bestModel
tree_model


    

    
Out[74]:





PipelineModel_4224923515634da2a63b

In [75]:

    

# Make predictions.
predictions = tree_model.transform(data)


    

In [76]:

    

# Select example rows to display.
predictions.select("features","label","predictedLabel").show(5)


    

    




+--------------------+-----+--------------+
|            features|label|predictedLabel|
+--------------------+-----+--------------+
|[7.4,0.7,0.0,1.9,...|    5|             5|
|[7.8,0.88,0.0,2.6...|    5|             5|
|[7.8,0.76,0.04,2....|    5|             5|
|[11.2,0.28,0.56,1...|    6|             6|
|[7.4,0.7,0.0,1.9,...|    5|             5|
+--------------------+-----+--------------+
only showing top 5 rows

In [ ]: