Standard scaler batch predict-Platform For AI(PAI)-阿里云帮助中心

Standard scaler batch predict applies standardization to batch data using a model trained by the Standard scaler train component. The algorithm uses the mean and variance to standardize the data, mapping values from different columns to a common range. This improves the stability and accuracy of downstream model predictions — particularly for large datasets where column ranges vary significantly.

This component assumes the input data follows a normal distribution.

Limits

Supported computing engines: MaxCompute and Flink.

Configure the component

Input ports

Input port (left to right)	Accepted data types	Recommended upstream component	Required
Input model for prediction	None	Standard scaler train	Yes
Input data for prediction	Numeric	Read table, Read CSV file	Yes

Parameters

Tab	Parameter	Description
Parameters	Output column names	Optional. By default, the prediction result columns replace the original input columns. The number of output columns must match the number of columns selected during training. Separate multiple column names with commas (,).
Parameters	Number of threads	The default value is 1.
Execution tuning	Number of Workers	Use with the Memory per worker parameter. Must be an integer from 1 to 9,999.
Execution tuning	Memory per worker (MB)	Must be from 1,024 MB to 65,536 MB.

Output ports

Output port (left to right)	Storage location	Recommended downstream component	Model type
Output result	No configuration required.	None	None

Example

The following PyAlink Script code performs the same operation as this component. sources[0] is the trained model and sources[1] is the batch data to transform.

from pyalink.alink import *

def main(sources, sinks, parameter):
    model = sources[0]
    batchData = sources[1]
    predictor = StandardScalerPredictBatchOp()
    result = predictor.linkFrom(model, batchData)
    result.link(sinks[0])
    BatchOperator.execute()