KerasSequential回归预测 (KerasSequentialRegressorPredictBatchOp)

Java 类名：com.alibaba.alink.operator.batch.regression.KerasSequentialRegressorPredictBatchOp

Python 类名：KerasSequentialRegressorPredictBatchOp

功能介绍

与 KerasSequential 回归训练组件对应的预测组件。

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	✓
inferBatchSize	推理数据批大小	推理数据批大小	Integer			256
reservedCols	算法保留列名	算法保留列	String[]			null

代码示例

** 以下代码仅用于示意，可能需要修改部分代码或者配置环境后才能正常运行！**

Python 代码

source = CsvSourceBatchOp() \
    .setFilePath("https://alink-release.oss-cn-beijing.aliyuncs.com/data-files/random_tensor.csv") \
    .setSchemaStr("tensor string, label double")

source = ToTensorBatchOp() \
    .setSelectedCol("tensor") \
    .setTensorDataType("DOUBLE") \
    .setTensorShape([200, 3]) \
    .linkFrom(source)

trainBatchOp = KerasSequentialRegressorTrainBatchOp() \
    .setTensorCol("tensor") \
    .setLabelCol("label") \
    .setLayers([
        "Conv1D(256, 5, padding='same', activation='relu')",
        "Conv1D(128, 5, padding='same', activation='relu')",
        "Dropout(0.1)",
        "MaxPooling1D(pool_size=8)",
        "Conv1D(128, 5, padding='same', activation='relu')",
        "Conv1D(128, 5, padding='same', activation='relu')",
        "Flatten()"
    ]) \
    .setOptimizer("Adam()") \
    .setNumEpochs(1) \
    .linkFrom(source)

predictBatchOp = KerasSequentialRegressorPredictBatchOp() \
    .setPredictionCol("pred") \
    .setReservedCols(["label"]) \
    .linkFrom(trainBatchOp, source)
predictBatchOp.lazyPrint(10)
BatchOperator.execute()

Java 代码

import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.dataproc.ToTensorBatchOp;
import com.alibaba.alink.operator.batch.regression.KerasSequentialRegressorPredictBatchOp;
import com.alibaba.alink.operator.batch.regression.KerasSequentialRegressorTrainBatchOp;
import com.alibaba.alink.operator.batch.source.CsvSourceBatchOp;
import org.junit.Test;

public class KerasSequentialRegressorTrainBatchOpTest {

	@Test
	public void testKerasSequentialRegressorTrainBatchOp() throws Exception {
		BatchOperator<?> source = new CsvSourceBatchOp()
			.setFilePath("https://alink-release.oss-cn-beijing.aliyuncs.com/data-files/random_tensor.csv")
			.setSchemaStr("tensor string, label double");

		source = new ToTensorBatchOp()
			.setSelectedCol("tensor")
			.setTensorDataType("DOUBLE")
			.setTensorShape(200, 3)
			.linkFrom(source);

		KerasSequentialRegressorTrainBatchOp trainBatchOp = new KerasSequentialRegressorTrainBatchOp()
			.setTensorCol("tensor")
			.setLabelCol("label")
			.setLayers(new String[] {
				"Conv1D(256, 5, padding='same', activation='relu')",
				"Conv1D(128, 5, padding='same', activation='relu')",
				"Dropout(0.1)",
				"MaxPooling1D(pool_size=8)",
				"Conv1D(128, 5, padding='same', activation='relu')",
				"Conv1D(128, 5, padding='same', activation='relu')",
				"Flatten()"
			})
			.setOptimizer("Adam()")
			.setNumEpochs(1)
			.linkFrom(source);

		KerasSequentialRegressorPredictBatchOp predictBatchOp = new KerasSequentialRegressorPredictBatchOp()
			.setPredictionCol("pred")
			.setReservedCols("label")
			.linkFrom(trainBatchOp, source);
		predictBatchOp.lazyPrint(10);
		BatchOperator.execute();
	}
}

运行结果

label	pred
1.0000	0.4822
0.0000	0.4826
0.0000	0.4752
0.0000	0.4702
1.0000	0.4907
1.0000	0.4992
0.0000	0.4866
1.0000	0.5045
0.0000	0.4994
1.0000	0.4837