Ridge Regression Training

Limits

The supported computing engines are MaxCompute, Flink, or DLC.

Algorithm principle

Ridge Regression is a biased estimation regression method used for the analysis of collinear data. It is an improved version of the least-squares estimation method. It sacrifices the unbiasedness of the least-squares method to obtain more practical and reliable regression coefficients. This trade-off comes at the cost of some information loss and reduced precision but provides a better fit for ill-conditioned data than the standard least-squares method.

Configure component parameters visually

• Input Port

  Port (from left to right)	Data type	Recommended upstream components	Required
  Data	None	Read Table Feature Engineering Data Preprocessing	Yes
  Model	None	Read Table	No

• Parameters

  Tab	Parameter	Description
  Field Settings	Target column name	The name of the target column in the input table.
  	Feature column array	This parameter cannot be configured if you have specified Vector column name. The names of the feature columns used for training. Note Feature column array and Vector column name are mutually exclusive. You can use only one of them to specify the input features for the algorithm.
  	Vector column name	This parameter cannot be configured if you have specified Feature column array. The name of the vector column. Note Feature column array and Vector column name are mutually exclusive. You can use only one of them to specify the input features for the algorithm.
  	Weight column name	The name of the weight column.
  Parameter Settings	Penalty factor: lambda	The coefficient of the regularization term. The data type is DOUBLE.
  	Convergence threshold	The threshold to determine whether the iterative method has converged. Default value: 1.0E-6.
  	Learning rate	Controls the speed at which parameters are updated during model training. Default value: 0.1.
  	Maximum number of iterations	The maximum number of iterations. Default value: 100.
  	Optimization method	The optimization method used to solve the problem. Valid values: LBFGS GD Newton SGD OWLQN
  Execution Tuning	Number of workers	Used with the Memory per worker parameter. This parameter must be a positive integer from 1 to 9999.
  	Memory per worker (MB)	The value ranges from 1024 MB to 64 × 1024 MB.

• Output ports

  Port (from left to right)	Data type	Downstream components
  Model	Regression model	Ridge Regression Prediction
  Model information	None	None
  Feature importance	None	None
  Linear model weight coefficients	None	None

Configure the component using code

from pyalink.alink import *

def main(sources, sinks, parameter):
    batchData = sources[0]
    ridge = RidgeRegTrainBatchOp()\
        .setLambda(0.1)\
        .setFeatureCols(["f0","f1"])\
        .setLabelCol("label")
    model = batchData.link(ridge)
    model.link(sinks[0])
    BatchOperator.execute()