Java 类名：com.alibaba.alink.operator.batch.similarity.VectorApproxNearestNeighborTrainBatchOp
Python 类名：VectorApproxNearestNeighborTrainBatchOp

功能介绍

该功能由训练和预测组成，训练时指定距离计算方式，生成最近邻模型
可选择的距离计算方式包含EUCLIDEAN和JACCARD两种，同时支持KDTREE和LSH两种近似方法。

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 | | —- | —- | —- | —- | —- | —- | —- |

| idCol | id列名 | id列名 | String | ✓ | | |

| selectedCol | 选中的列名 | 计算列对应的列名 | String | ✓ | 所选列类型为 [DENSE_VECTOR, SPARSE_VECTOR, STRING, VECTOR] | |

| metric | 距离度量方式 | 距离类型 | String | | “EUCLIDEAN”, “JACCARD” | “EUCLIDEAN” |

| numHashTables | 哈希表的数目 | 哈希表的数目 | Integer | | | 1 |

| numProjectionsPerTable | 每个哈希表中的哈希函数个数 | 每个哈希表中的哈希函数个数 | Integer | | | 1 |

| projectionWidth | 桶的宽度 | 桶的宽度 | Double | | | 1.0 |

| seed | 采样种子 | 采样种子 | Long | | | 0 |

| solver | 近似方法 | 近似方法，包括KDTREE和LSH | String | | “KDTREE”, “LSH” | “KDTREE” |

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"],
  9.     [1, "1 1 1"],
  10.     [2, "2 2 2"]
  11. ])
  13. inOp = BatchOperator.fromDataframe(df, schemaStr='id int, vec string')
  14. train = VectorApproxNearestNeighborTrainBatchOp().setIdCol("id").setSelectedCol("vec").linkFrom(inOp)
  15. predict = VectorApproxNearestNeighborPredictBatchOp().setSelectedCol("vec").setTopN(3).linkFrom(train, inOp)
  16. predict.print()

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.similarity.VectorApproxNearestNeighborPredictBatchOp;
  5. import com.alibaba.alink.operator.batch.similarity.VectorApproxNearestNeighborTrainBatchOp;
  6. import com.alibaba.alink.operator.batch.source.MemSourceBatchOp;
  7. import org.junit.Test;
  9. import java.util.Arrays;
  10. import java.util.List;
  12. public class VectorApproxNearestNeighborTrainBatchOpTest {
  13.     @Test
  14.     public void testVectorApproxNearestNeighborTrainBatchOp() throws Exception {
  15.         List <Row> df = Arrays.asList(
  16.             Row.of(0, "0 0 0"),
  17.             Row.of(1, "1 1 1"),
  18.             Row.of(2, "2 2 2")
  19.         );
  20.         BatchOperator <?> inOp = new MemSourceBatchOp(df, "id int, vec string");
  21.         BatchOperator <?> train = new VectorApproxNearestNeighborTrainBatchOp().setIdCol("id").setSelectedCol("vec")
  22.             .linkFrom(inOp);
  23.         BatchOperator <?> predict = new VectorApproxNearestNeighborPredictBatchOp().setSelectedCol("vec").setTopN(3)
  24.             .linkFrom(train, inOp);
  25.         predict.print();
  26.     }
  27. }

运行结果

| id | vec | | —- | —- |

| 0 | {“ID”:”[0,1,2]”,”METRIC”:”[0.0,1.7320508075688772,3.4641016151377544]”} |

| 1 | {“ID”:”[1,2,0]”,”METRIC”:”[0.0,1.7320508075688772,1.7320508075688772]”} |

| 2 | {“ID”:”[2,1,0]”,”METRIC”:”[0.0,1.7320508075688772,3.4641016151377544]”} |