ALINK(二十)：数据处理（六）数值型数据处理（二）标准化 (StandardScalerPredictBatchOp/StandardScalerTrainBatchOp )

标准化训练 (StandardScalerTrainBatchOp)

Java 类名：com.alibaba.alink.operator.batch.dataproc.StandardScalerTrainBatchOp

Python 类名：StandardScalerTrainBatchOp

功能介绍

标准化是对数据进行按正态化处理的组件

训练过程计算数据的均值和标准差，在预测组件中使用模型结果

参数说明

名称	中文名称	描述	类型	是否必须？	默认值
selectedCols	选择的列名	计算列对应的列名列表	String[]	✓
withMean	是否使用均值	是否使用均值，默认使用	Boolean		true
withStd	是否使用标准差	是否使用标准差，默认使用	Boolean		true

代码示例

Python 代码

from pyalink.alink import *
import pandas as pd
useLocalEnv(1)
df = pd.DataFrame([
            ["a", 10.0, 100],
            ["b", -2.5, 9],
            ["c", 100.2, 1],
            ["d", -99.9, 100],
            ["a", 1.4, 1],
            ["b", -2.2, 9],
            ["c", 100.9, 1]
])
             
colnames = ["col1", "col2", "col3"]
selectedColNames = ["col2", "col3"]
inOp = BatchOperator.fromDataframe(df, schemaStr='col1 string, col2 double, col3 long')
         
# train
trainOp = StandardScalerTrainBatchOp()
           .setSelectedCols(selectedColNames)
trainOp.linkFrom(inOp)
# batch predict
predictOp = StandardScalerPredictBatchOp()
predictOp.linkFrom(trainOp, inOp).print()

Java 代码

import org.apache.flink.types.Row;
import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.dataproc.StandardScalerPredictBatchOp;
import com.alibaba.alink.operator.batch.dataproc.StandardScalerTrainBatchOp;
import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
import org.junit.Test;
import java.util.Arrays;
import java.util.List;
public class StandardScalerTrainBatchOpTest {
  @Test
  public void testStandardScalerTrainBatchOp() throws Exception {
    List <Row> df = Arrays.asList(
      Row.of("a", 10.0, 100),
      Row.of("b", -2.5, 9),
      Row.of("c", 100.2, 1),
      Row.of("d", -99.9, 100),
      Row.of("a", 1.4, 1),
      Row.of("b", -2.2, 9),
      Row.of("c", 100.9, 1)
    );
    String[] selectedColNames = new String[] {"col2", "col3"};
    BatchOperator <?> inOp = new MemSourceBatchOp(df, "col1 string, col2 double, col3 int");
    BatchOperator <?> trainOp = new StandardScalerTrainBatchOp()
      .setSelectedCols(selectedColNames);
    trainOp.linkFrom(inOp);
    BatchOperator <?> predictOp = new StandardScalerPredictBatchOp();
    predictOp.linkFrom(trainOp, inOp).print();
  }
}

运行结果

col1	col2	col3
a	-0.0784	1.4596
b	-0.2592	-0.4814
c	1.2270	-0.6521
d	-1.6687	1.4596
a	-0.2028	-0.6521
b	-0.2549	-0.4814
c	1.2371	-0.6521

 

 

 

 

 

 

 

标准化批预测 (StandardScalerPredictBatchOp)

Java 类名：com.alibaba.alink.operator.batch.dataproc.StandardScalerPredictBatchOp

Python 类名：StandardScalerPredictBatchOp

功能介绍

标准化是对数据进行按正态化处理的组件

使用标准化训练组件训练的模型，对数据做标准化处理

参数说明

名称	中文名称	描述	类型	是否必须？	默认值
outputCols	输出结果列列名数组	输出结果列列名数组，可选，默认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([
            ["a", 10.0, 100],
            ["b", -2.5, 9],
            ["c", 100.2, 1],
            ["d", -99.9, 100],
            ["a", 1.4, 1],
            ["b", -2.2, 9],
            ["c", 100.9, 1]
])
             
colnames = ["col1", "col2", "col3"]
selectedColNames = ["col2", "col3"]
inOp = BatchOperator.fromDataframe(df, schemaStr='col1 string, col2 double, col3 long')
         
# train
trainOp = StandardScalerTrainBatchOp()
           .setSelectedCols(selectedColNames)
trainOp.linkFrom(inOp)
# batch predict
predictOp = StandardScalerPredictBatchOp()
predictOp.linkFrom(trainOp, inOp).print()

Java 代码

import org.apache.flink.types.Row;
import com.alibaba.alink.operator.batch.BatchOperator;
import com.alibaba.alink.operator.batch.dataproc.StandardScalerPredictBatchOp;
import com.alibaba.alink.operator.batch.dataproc.StandardScalerTrainBatchOp;
import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
import org.junit.Test;
import java.util.Arrays;
import java.util.List;
public class StandardScalerPredictBatchOpTest {
  @Test
  public void testStandardScalerPredictBatchOp() throws Exception {
    List <Row> df = Arrays.asList(
      Row.of("a", 10.0, 100),
      Row.of("b", -2.5, 9),
      Row.of("c", 100.2, 1),
      Row.of("d", -99.9, 100),
      Row.of("a", 1.4, 1),
      Row.of("b", -2.2, 9),
      Row.of("c", 100.9, 1)
    );
    String[] selectedColNames = new String[] {"col2", "col3"};
    BatchOperator <?> inOp = new MemSourceBatchOp(df, "col1 string, col2 double, col3 int");
    BatchOperator <?> trainOp = new StandardScalerTrainBatchOp()
      .setSelectedCols(selectedColNames);
    trainOp.linkFrom(inOp);
    BatchOperator <?> predictOp = new StandardScalerPredictBatchOp();
    predictOp.linkFrom(trainOp, inOp).print();
  }
}

运行结果

col1	col2	col3
a	-0.0784	1.4596
b	-0.2592	-0.4814
c	1.2270	-0.6521
d	-1.6687	1.4596
a	-0.2028	-0.6521
b	-0.2549	-0.4814
c	1.2371	-0.6521