Java 类名：com.alibaba.alink.operator.batch.feature.PcaPredictBatchOp
Python 类名：PcaPredictBatchOp

功能介绍

主成分分析，是考察多个变量间相关性一种多元统计方法，研究如何通过少数几个主成分来揭示多个变量间的内部结构，即从原始变量中导出少数几个主成分，使它们尽可能多地保留原始变量的信息，且彼此间互不相关，作为新的综合指标。详细介绍请见维基百科链接wiki。
主成分分析功能包含主成分分析训练和主成分分析预测（批和流)。

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 所选列类型为 [DENSE_VECTOR, SPARSE_VECTOR, STRING, VECTOR] null
numThreads 组件多线程线程个数 组件多线程线程个数 Integer 1

代码示例

Python 代码

  1. from pyalink.alink import *
  3. import pandas as pd
  5. useLocalEnv(1)
  7. df = pd.DataFrame([
  8.         [0.0,0.0,0.0],
  9.         [0.1,0.2,0.1],
  10.         [0.2,0.2,0.8],
  11.         [9.0,9.5,9.7],
  12.         [9.1,9.1,9.6],
  13.         [9.2,9.3,9.9]
  14. ])
  16. # batch source 
  17. inOp = BatchOperator.fromDataframe(df, schemaStr='x1 double, x2 double, x3 double')
  19. trainOp = PcaTrainBatchOp()\
  20.        .setK(2)\
  21.        .setSelectedCols(["x1","x2","x3"])
  23. predictOp = PcaPredictBatchOp()\
  24.         .setPredictionCol("pred")
  26. # batch train
  27. inOp.link(trainOp)
  29. # batch predict
  30. predictOp.linkFrom(trainOp,inOp)
  32. predictOp.print()
  34. # stream predict
  35. inStreamOp = StreamOperator.fromDataframe(df, schemaStr='x1 double, x2 double, x3 double')
  37. predictStreamOp = PcaPredictStreamOp(trainOp)\
  38.         .setPredictionCol("pred")
  40. predictStreamOp.linkFrom(inStreamOp)
  42. predictStreamOp.print()
  44. StreamOperator.execute()

Java 代码

  1. import org.apache.flink.types.Row;
  3. import com.alibaba.alink.operator.batch.BatchOperator;
  4. import com.alibaba.alink.operator.batch.feature.PcaPredictBatchOp;
  5. import com.alibaba.alink.operator.batch.feature.PcaTrainBatchOp;
  6. import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
  7. import com.alibaba.alink.operator.stream.StreamOperator;
  8. import com.alibaba.alink.operator.stream.feature.PcaPredictStreamOp;
  9. import com.alibaba.alink.operator.stream.source.MemSourceStreamOp;
  10. import org.junit.Test;
  12. import java.util.Arrays;
  13. import java.util.List;
  15. public class PcaPredictBatchOpTest {
  16.     @Test
  17.     public void testPcaPredictBatchOp() throws Exception {
  18.         List <Row> df = Arrays.asList(
  19.             Row.of(0.0, 0.0, 0.0),
  20.             Row.of(0.1, 0.2, 0.1),
  21.             Row.of(0.2, 0.2, 0.8),
  22.             Row.of(9.0, 9.5, 9.7),
  23.             Row.of(9.1, 9.1, 9.6),
  24.             Row.of(9.2, 9.3, 9.9)
  25.         );
  26.         BatchOperator <?> inOp = new MemSourceBatchOp(df, "x1 double, x2 double, x3 double");
  27.         BatchOperator <?> trainOp = new PcaTrainBatchOp()
  28.             .setK(2)
  29.             .setSelectedCols("x1", "x2", "x3");
  30.         BatchOperator <?> predictOp = new PcaPredictBatchOp()
  31.             .setPredictionCol("pred");
  32.         inOp.link(trainOp);
  33.         predictOp.linkFrom(trainOp, inOp);
  34.         predictOp.print();
  35.         StreamOperator <?> inStreamOp = new MemSourceStreamOp(df, "x1 double, x2 double, x3 double");
  36.         StreamOperator <?> predictStreamOp = new PcaPredictStreamOp(trainOp)
  37.             .setPredictionCol("pred");
  38.         predictStreamOp.linkFrom(inStreamOp);
  39.         predictStreamOp.print();
  40.         StreamOperator.execute();
  41.     }
  42. }

运行结果

| x1 | x2 | x3 | pred | | —- | —- | —- | —- |

| 9.0 | 9.5 | 9.7 | 3.2280384305400736,1.1516225426477789E-4 |

| 0.2 | 0.2 | 0.8 | 0.13565076707329407,0.09003329494282108 |

| 9.2 | 9.3 | 9.9 | 3.250783163664603,0.0456526246528135 |

| 9.1 | 9.1 | 9.6 | 3.182618319978973,0.027469531992220464 |

| 0.1 | 0.2 | 0.1 | 0.045855205015063565,-0.012182917696915518 |

| 0.0 | 0.0 | 0.0 | 0.0,0.0 |