Decision Tree

Setting Up a Decision Tree Classifier

Note: Make sure you have your training and test data already vectorized and ready to go before you begin trying to fit the machine learning model to unprepped data.

Load in required libraries

from pyspark.ml.classification import DecisionTreeClassifier
from pyspark.ml.tuning import ParamGridBuilder, CrossValidator
from pyspark.ml.evaluation import BinaryClassificationEvaluator

Initialize Decision Tree object

dt = DecisionTreeClassifier(labelCol="label", featuresCol="features")

Create a parameter grid for tuning the model

dtparamGrid = (ParamGridBuilder()
             .addGrid(dt.maxDepth, [2, 5, 10])
             .addGrid(dt.maxBins, [10, 20])
             .build())

Define how you want the model to be evaluated

dtevaluator = BinaryClassificationEvaluator(rawPredictionCol="rawPrediction")

Define the type of cross-validation you want to perform

# Create 5-fold CrossValidator
dtcv = CrossValidator(estimator = dt,
                      estimatorParamMaps = dtparamGrid,
                      evaluator = dtevaluator,
                      numFolds = 5)

Fit the model to the data

dtcvModel = dtcv.fit(train)
print(dtcvModel)

Score the testing dataset using your fitted model for evaluation purposes

dtpredictions = dtcvModel.transform(test)

Evaluate the model

print('Accuracy:', dtevaluator.evaluate(dtpredictions))
print('AUC:', BinaryClassificationMetrics(dtpredictions['label','prediction'].rdd).areaUnderROC)
print('PR:', BinaryClassificationMetrics(dtpredictions['label','prediction'].rdd).areaUnderPR)

Note: When you use the CrossValidator function to set up cross-validation of your models, the resulting model object will have all the runs included, but will only use the best model when you interact with the model object using other functions like evaluate or transform.