Spark随机森林实现票房预测

public void predict() throws IOException{
  SparkConf conf = new SparkConf().setAppName("SVM").setMaster("local");
  conf.set("spark.testing.memory", "2147480000");
  SparkContext sc = new SparkContext(conf);
  SQLContext sqlContext = new SQLContext(sc);


  // Load and parse the data file, converting it to a DataFrame.
  DataFrame trainData = sqlContext.read().format("libsvm").load(this.trainFile);
  DataFrame testData = sqlContext.read().format("libsvm").load(this.testFile);

  // Index labels, adding metadata to the label column.
  // Fit on whole dataset to include all labels in index.
  StringIndexerModel labelIndexer = new StringIndexer()
   .setInputCol("label")
   .setOutputCol("indexedLabel")
   .fit(trainData);
  // Automatically identify categorical features, and index them.
  // Set maxCategories so features with > 4 distinct values are treated as continuous.
  VectorIndexerModel featureIndexer = new VectorIndexer()
   .setInputCol("features")
   .setOutputCol("indexedFeatures")
   .setMaxCategories(4)
   .fit(trainData);

  // Split the data into training and test sets (30% held out for testing)
//  DataFrame[] splits = trainData.randomSplit(new double[] {0.9, 0.1});
//  trainData = splits[0];
//  testData = splits[1];

  // Train a RandomForest model.
  RandomForestClassifier rf = new RandomForestClassifier()
   .setLabelCol("indexedLabel")
   .setFeaturesCol("indexedFeatures")
   .setNumTrees(20);

  // Convert indexed labels back to original labels.
  IndexToString labelConverter = new IndexToString()
   .setInputCol("prediction")
   .setOutputCol("predictedLabel")
   .setLabels(labelIndexer.labels());

  // Chain indexers and forest in a Pipeline
  Pipeline pipeline = new Pipeline()
   .setStages(new PipelineStage[] {labelIndexer, featureIndexer, rf, labelConverter});

  // Train model. This also runs the indexers.
  PipelineModel model = pipeline.fit(trainData);

  // Make predictions.
  DataFrame predictions = model.transform(testData);

  // Select example rows to display.
  predictions.select("predictedLabel", "label", "features").show(200);

  // Select (prediction, true label) and compute test error
  MulticlassClassificationEvaluator evaluator = new MulticlassClassificationEvaluator()
   .setLabelCol("indexedLabel")
   .setPredictionCol("prediction")
   .setMetricName("precision");
  double accuracy = evaluator.evaluate(predictions);
  System.out.println("Test Error = " + (1.0 - accuracy));

  RandomForestClassificationModel rfModel = (RandomForestClassificationModel)(model.stages()[2]);
//  System.out.println("Learned classification forest model:\n" + rfModel.toDebugString());

  DataFrame resultDF = predictions.select("predictedLabel");
  JavaRDD<Row> resultRow = resultDF.toJavaRDD();
  JavaRDD<String> result = resultRow.map(new Result());
  this.resultList = result.collect();
  for(String one: resultList){
   System.out.println(one);
  }
 }

Tree 16 (weight 1.0):
 If (feature 10 in {0.0})
  If (feature 48 <= 110.0)
  If (feature 86 <= 13698.87)
  If (feature 21 in {0.0})
  If (feature 54 in {0.0})
   Predict: 0.0
  Else (feature 54 not in {0.0})
   Predict: 1.0
  Else (feature 21 not in {0.0})
  Predict: 0.0
  Else (feature 86 > 13698.87)
  If (feature 21 in {0.0})
  If (feature 85 <= 39646.9)
   Predict: 2.0
  Else (feature 85 > 39646.9)
   Predict: 3.0
  Else (feature 21 not in {0.0})
  Predict: 3.0
  Else (feature 48 > 110.0)
  If (feature 85 <= 15003.3)
  If (feature 9 in {0.0})
  If (feature 54 in {0.0})
   Predict: 0.0
  Else (feature 54 not in {0.0})
   Predict: 2.0
  Else (feature 9 not in {0.0})
  Predict: 2.0
  Else (feature 85 > 15003.3)
  If (feature 65 in {0.0})
  If (feature 85 <= 66065.0)
   Predict: 3.0
  Else (feature 85 > 66065.0)
   Predict: 2.0
  Else (feature 65 not in {0.0})
  Predict: 3.0
 Else (feature 10 not in {0.0})
  If (feature 51 in {0.0})
  If (feature 85 <= 6958.4)
  If (feature 11 in {0.0})
  If (feature 50 <= 1.0)
   Predict: 1.0
  Else (feature 50 > 1.0)
   Predict: 0.0
  Else (feature 11 not in {0.0})
  Predict: 0.0
  Else (feature 85 > 6958.4)
  If (feature 5 in {0.0})
  If (feature 4 in {0.0})
   Predict: 3.0
  Else (feature 4 not in {0.0})
   Predict: 1.0
  Else (feature 5 not in {0.0})
  Predict: 2.0
  Else (feature 51 not in {0.0})
  If (feature 48 <= 148.0)
  If (feature 0 in {0.0})
  If (feature 6 in {0.0})
   Predict: 2.0
  Else (feature 6 not in {0.0})
   Predict: 0.0
  Else (feature 0 not in {0.0})
  If (feature 50 <= 4.0)
   Predict: 2.0
  Else (feature 50 > 4.0)
   Predict: 3.0
  Else (feature 48 > 148.0)
  If (feature 9 in {0.0})
  If (feature 49 <= 3.0)
   Predict: 2.0
  Else (feature 49 > 3.0)
   Predict: 0.0
  Else (feature 9 not in {0.0})
  If (feature 36 in {0.0})
   Predict: 3.0
  Else (feature 36 not in {0.0})
   Predict: 2.0