GBDT排序预测 (GbdtRegPredictStreamOp)

Java 类名:com.alibaba.alink.operator.stream.regression.GbdtRegPredictStreamOp

Python 类名:GbdtRegPredictStreamOp

功能介绍

  • gbdt(Gradient Boosting Decision Trees)回归,是经典的基于boosting的有监督学习模型,可以用来解决回归问题

  • 支持连续特征和离散特征

  • 支持数据采样和特征采样

参数说明

名称 中文名称 描述 类型 是否必须? 取值范围 默认值
predictionCol 预测结果列名 预测结果列名 String
modelFilePath 模型的文件路径 模型的文件路径 String null
reservedCols 算法保留列名 算法保留列 String[] null
vectorCol 向量列名 向量列对应的列名,默认值是null String null
numThreads 组件多线程线程个数 组件多线程线程个数 Integer 1
modelStreamFilePath 模型流的文件路径 模型流的文件路径 String null
modelStreamScanInterval 扫描模型路径的时间间隔 描模型路径的时间间隔,单位秒 Integer 10
modelStreamStartTime 模型流的起始时间 模型流的起始时间。默认从当前时刻开始读。使用yyyy-mm-dd hh:mm:ss.fffffffff格式,详见Timestamp.valueOf(String s) String null

代码示例

Python 代码

from pyalink.alink import *

import pandas as pd

useLocalEnv(1)

df = pd.DataFrame([
    [1.0, "A", 0, 0, 0],
    [2.0, "B", 1, 1, 0],
    [3.0, "C", 2, 2, 1],
    [4.0, "D", 3, 3, 1]
])
batchSource = BatchOperator.fromDataframe(
    df, schemaStr='f0 double, f1 string, f2 int, f3 int, label int')
streamSource = StreamOperator.fromDataframe(
    df, schemaStr='f0 double, f1 string, f2 int, f3 int, label int')

trainOp = GbdtRegTrainBatchOp()\
    .setLearningRate(1.0)\
    .setNumTrees(3)\
    .setMinSamplesPerLeaf(1)\
    .setLabelCol('label')\
    .setFeatureCols(['f0', 'f1', 'f2', 'f3'])\
    .linkFrom(batchSource)
predictBatchOp = GbdtRegPredictBatchOp()\
    .setPredictionCol('pred')
predictStreamOp = GbdtRegPredictStreamOp(trainOp)\
    .setPredictionCol('pred')

predictBatchOp.linkFrom(trainOp, batchSource).print()
predictStreamOp.linkFrom(streamSource).print()

StreamOperator.execute()

Java 代码

import org.apache.flink.types.Row;

import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.regression.GbdtRegPredictBatchOp;
import com.alibaba.alink.operator.batch.regression.GbdtRegTrainBatchOp;
import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
import com.alibaba.alink.operator.stream.StreamOperator;
import com.alibaba.alink.operator.stream.regression.GbdtRegPredictStreamOp;
import com.alibaba.alink.operator.stream.source.MemSourceStreamOp;
import org.junit.Test;

import java.util.Arrays;
import java.util.List;

public class GbdtRegPredictStreamOpTest {

	@Test
	public void testGbdtRegPredictStreamOp() throws Exception {
		List <Row> df = Arrays.asList(
			Row.of(1.0, "A", 0, 0, 0),
			Row.of(2.0, "B", 1, 1, 0),
			Row.of(3.0, "C", 2, 2, 1),
			Row.of(4.0, "D", 3, 3, 1)
		);
		BatchOperator <?> batchSource = new MemSourceBatchOp(
			df, "f0 double, f1 string, f2 int, f3 int, label int");
		StreamOperator <?> streamSource = new MemSourceStreamOp(
			df, "f0 double, f1 string, f2 int, f3 int, label int");
		BatchOperator <?> trainOp = new GbdtRegTrainBatchOp()
			.setLearningRate(1.0)
			.setNumTrees(3)
			.setMinSamplesPerLeaf(1)
			.setLabelCol("label")
			.setFeatureCols("f0", "f1", "f2", "f3")
			.linkFrom(batchSource);
		BatchOperator <?> predictBatchOp = new GbdtRegPredictBatchOp()
			.setPredictionCol("pred");
		StreamOperator <?> predictStreamOp = new GbdtRegPredictStreamOp(trainOp)
			.setPredictionCol("pred");
		predictBatchOp.linkFrom(trainOp, batchSource).print();
		predictStreamOp.linkFrom(streamSource).print();
		StreamOperator.execute();
	}
}

运行结果

批预测结果

f0 f1 f2 f3 label pred
1.0000 A 0 0 0 0.0000
2.0000 B 1 1 0 0.0000
3.0000 C 2 2 1 1.0000
4.0000 D 3 3 1 1.0000

流预测结果

f0 f1 f2 f3 label pred
2.0000 B 1 1 0 0.0000
4.0000 D 3 3 1 1.0000
1.0000 A 0 0 0 0.0000
3.0000 C 2 2 1 1.0000