Description

MinMaxScaler transforms a dataSet of rows, rescaling each feature to a specific range [min, max). (often [0, 1]). MinMaxScalerTrain will train a model.

Parameters

Name Description Type Required？ Default Value
selectedCol Name of the selected column used for processing String ✓
min Lower bound after transformation. Double 0.0
max Upper bound after transformation. Double 1.0

Script Example

Script

  1. data = np.array([["a", "10.0, 100"],\
  2.     ["b", "-2.5, 9"],\
  3.     ["c", "100.2, 1"],\
  4.     ["d", "-99.9, 100"],\
  5.     ["a", "1.4, 1"],\
  6.     ["b", "-2.2, 9"],\
  7.     ["c", "100.9, 1"]])
  8. df = pd.DataFrame({"col" : data[:,0], "vec" : data[:,1]})
  9. data = dataframeToOperator(df, schemaStr="col string, vec string",op_type="batch")
  10. trainOp = VectorMinMaxScalerTrainBatchOp()\
  11.            .setSelectedCol("vec")
  12. model = trainOp.linkFrom(data) 
  14. batchPredictOp = VectorMinMaxScalerPredictBatchOp()
  15. batchPredictOp.linkFrom(model, data).collectToDataframe()

Result

col1 vec
a 0.5473107569721115,1.0
b 0.4850597609561753,0.08080808080808081
c 0.9965139442231076,0.0
d 0.0,1.0
a 0.5044820717131474,0.0
b 0.4865537848605578,0.08080808080808081
c 1.0,0.0