Weighted sampling-Platform For AI(PAI)-阿里云帮助中心

This component generates sample data using weighted sampling. The weight column must be of the DOUBLE or BIGINT data type. Records are sampled based on the values in this column. For example, a record with a weight of 1.2 has a higher probability of being sampled than a record with a weight of 1.0.

Configure the component

You can configure the parameters for the Weighted Sampling component in one of the following ways.

Method 1: Use the GUI

You can configure the component parameters on the workflow page in Designer.

Tab	Parameter	Description
Parameters setting	Sample size	A positive integer.
	Sampling ratio	A floating-point number in the range of (0, 1).
	Sampling with replacement	Select this option to enable sampling with replacement. By default, sampling is performed without replacement.
	Weight column	Select a column to use for weighting. The column must be of the DOUBLE or BIGINT type. Each value represents the weight of the corresponding record. Normalization is not required.
	Random seed	The system generates this by default.
Execution tuning	Number of cores	The value must be a positive integer. The system automatically assigns a default value.
Execution tuning	Kernel Memory Allocation	A positive integer in the range of (1, 65536). The system automatically allocates the memory by default.

Method 2: Use a PAI command

You can configure the component parameters using a PAI command in the SQL script component. For more information, see SQL script.

PAI -name WeightedSample
    -project algo_public
    -Dlifecycle="28"
    -DoutputTableName="test2"
    -DprobCol="previous"
    -Dreplace="false"
    -DsampleSize="500"
    -DinputPartitions="pt=20150501"
    -DinputTableName="bank_data_partition";

Parameter Name	Required	Description	Default value
inputTableName	Yes	The name of the input table.	None
inputTablePartitions	No	The partitions in the input table to use for training. The following formats are supported: Partition_name=value name1=value1/name2=value2: a multi-level format Note If you specify multiple partitions, separate them with commas (,).	All partitions
outputTableName	Yes	The output table.	None
sampleSize	No	The number of samples to generate. Note If both the sampleSize and sampleRatio parameters are left empty, an error is reported. If both the sampleSize and sampleRatio parameters are specified, the sampleSize parameter takes precedence.	None
sampleRatio	No	The sampling ratio. This parameter specifies a floating-point number in the range of (0, 1).	None
probCol	Yes	The column to use for weighting. Each value represents the weight of the corresponding record. Normalization is not required. The column must be of the DOUBLE or BIGINT type.	None
replace	No	Specifies whether to perform sampling with replacement. This parameter is of the BOOLEAN type.	false (Sampling without replacement is performed by default.)
randomSeed	No	The random seed. The value must be a positive integer.	The system automatically assigns a value.
lifecycle	No	The lifecycle of the output table in days. The value must be an integer in the range of [1, 3650].	None
coreNum	No	The number of cores for computing. The value must be a positive integer.	The system assigns this automatically.
memSizePerCore	No	The memory size for each core, in MB. The value must be an integer in the range of (1, 65536).	This is automatically assigned by the system.