GMM预测

功能介绍

基于GaussianMixture模型进行聚类预测。

参数说明

名称 中文名称 描述 类型 是否必须？ 默认值
vectorCol 向量列名 向量列对应的列名 String ✓
predictionCol 预测结果列名 预测结果列名 String ✓
predictionDetailCol 预测详细信息列名 预测详细信息列名 String
reservedCols 算法保留列名 算法保留列 String[] null

脚本示例

脚本代码

  1. data = np.array([
  2.     ["-0.6264538 0.1836433"],
  3.     ["-0.8356286 1.5952808"],
  4.     ["0.3295078 -0.8204684"],
  5.     ["0.4874291 0.7383247"],
  6.     ["0.5757814 -0.3053884"],
  7.     ["1.5117812 0.3898432"],
  8.     ["-0.6212406 -2.2146999"],
  9.     ["11.1249309 9.9550664"],
  10.     ["9.9838097 10.9438362"],
  11.     ["10.8212212 10.5939013"],
  12.     ["10.9189774 10.7821363"],
  13.     ["10.0745650 8.0106483"],
  14.     ["10.6198257 9.9438713"],
  15.     ["9.8442045 8.5292476"],
  16.     ["9.5218499 10.4179416"],
  17. ])
  19. df_data = pd.DataFrame({
  20.     "features": data[:, 0],
  21. })
  23. data = dataframeToOperator(df_data, schemaStr='features string', op_type='batch')
  25. gmm = GmmTrainBatchOp() \
  26.     .setVectorCol("features") \
  27.     .setTol(0.)
  29. model = gmm.linkFrom(data)
  31. predictor = GmmPredictBatchOp() \
  32.     .setPredictionCol("cluster_id") \
  33.     .setVectorCol("features") \
  34.     .setPredictionDetailCol("cluster_detail")
  36. predictor.linkFrom(model, data).print()

脚本运行结果

  1.                  features  cluster_id              cluster_detail
  2. 0    -0.6264538 0.1836433           0   1.0 4.275273913994647E-92
  3. 1    -0.8356286 1.5952808           0  1.0 1.0260377730322135E-92
  4. 2    0.3295078 -0.8204684           0  1.0 1.0970173367582936E-80
  5. 3     0.4874291 0.7383247           0    1.0 3.30217313232611E-75
  6. 4    0.5757814 -0.3053884           0   1.0 3.163811360527691E-76
  7. 5     1.5117812 0.3898432           0  1.0 2.1018052308786076E-62
  8. 6   -0.6212406 -2.2146999           0   1.0 6.772270268625197E-97
  9. 7    11.1249309 9.9550664           1  3.1567838012477083E-56 1.0
  10. 8    9.9838097 10.9438362           1  1.9024447346702333E-51 1.0
  11. 9   10.8212212 10.5939013           1  2.8009730987296404E-56 1.0
  12. 10  10.9189774 10.7821363           1  1.7209132744891575E-57 1.0
  13. 11   10.0745650 8.0106483           1   2.864269663513225E-43 1.0
  14. 12   10.6198257 9.9438713           1    5.77327399194046E-53 1.0
  15. 13    9.8442045 8.5292476           1  2.5273123050926845E-43 1.0
  16. 14   9.5218499 10.4179416           1  1.7314580596765865E-46 1.0