基于spark框架实现的随机森林例子

数据处理：

        Dataset<Row> dataset = datasets[0];
        Dataset<Row>[] splits = dataset.randomSplit(new double[]{0.7, 0.3});
        Dataset<Row> training = splits[0];
        Dataset<Row> test = splits[1];

---

调用ml中的模型：

        RandomForestClassifier rf = new RandomForestClassifier();
        RandomForestClassificationModel model = rf.fit(training);

---

设置参数：

        rf.setMaxBins(5);
        rf.setMaxDepth(10);
        rf.setImpurity("gini");
        rf.setNumTrees(5);
        rf.setSeed(12345);

---

 预测评估：

        MulticlassClassificationEvaluator evaluator = new MulticlassClassificationEvaluator();
        double accuracy = evaluator.evaluate(predictions);
        double error = 1-accuracy;

---

返回结果：

        List<Row> data = Arrays.asList(
                RowFactory.create(accuracy),
                RowFactory.create(error)
        );

        StructType schema = new StructType(new StructField[]{
                createStructField("accuracy", DoubleType, false),
                createStructField("error", DoubleType, false)
        });

        return this.sparkSession.createDataFrame(data,schema);