Scorecard Prediction component-Platform For AI(PAI)-阿里云帮助中心

Scorecard Prediction is a machine learning technique that applies a scorecard model to new data to predict future performance or risk. This model is typically generated by a scorecard training component. The Scorecard Prediction component uses this model to evaluate and score input data, which aids in decision-making and risk management.

Configure the component

Method 1: Use the UI

In Designer, add the Scorecard Prediction component to your pipeline and configure its parameters in the right-side pane.

Tab	Parameter	Description
Field Settings	Feature Columns	Select the raw feature columns to use for prediction. By default, all columns are selected.
	Reserved Columns	Select columns to append to the prediction result table without processing, such as ID columns and label columns.
	Output Variable Score	Specifies whether to output the score for each feature variable. The final prediction score is the sum of the score for the intercept term and all variable scores.
Tuning	Number of cores	The number of CPU cores to use. By default, the system assigns a value automatically.
Tuning	Memory per core	The memory allocated to each CPU core. By default, the system assigns a value automatically.

Method 2: Use a PAI command

Configure the Scorecard Prediction component parameters using a PAI command. You can run PAI commands in the SQL Script component. For more information, see SQL Script.

pai -name=lm_predict
    -project=algo_public
    -DinputFeatureTableName=input_data_table
    -DinputModelTableName=input_model_table
    -DmetaColNames=sample_key,label
    -DfeatureColNames=fea1,fea2
    -DoutputTableName=output_score_table

Parameter	Required	Default	Description
inputFeatureTableName	Yes	None	The input feature table.
inputFeatureTablePartitions	No	The entire table	The partitions to use from the input feature table.
inputModelTableName	Yes	None	The input model table.
featureColNames	No	All columns	The feature columns to use from the input table.
metaColNames	No	None	The columns that are passed through to the output table without being processed, such as a label and sample_id.
outputFeatureScore	No	false	Specifies whether to output variable scores in the prediction results. Valid values: true: Outputs the variable scores. false: Does not output the variable scores.
outputTableName	Yes	None	The output table for the prediction results.
lifecycle	No	None	The lifecycle of the output table.
coreNum	No	Automatically determined	The number of cores.
memSizePerCore	No	Automatically determined	The memory size per core, in MB.

Outputs

The following is a sample output from the Scorecard Prediction component. After scoring, the output table contains the original label column, such as churn (with a value of 0 or 1), and prediction columns. The churn column is a reserved column that is passed through to the output and is not part of the prediction results. The following table describes the prediction columns.

Name	Type	Description
prediction_score	DOUBLE	The predicted score. For a linear model, this is the sum of each feature value multiplied by its corresponding model weight. In a scorecard model, if score transformation is performed, this column contains the transformed score.
prediction_prob	DOUBLE	For binary classification, this is the predicted probability of the positive class. This value is calculated by applying a Sigmoid transformation to the original score, before any score transformation occurs.
prediction_detail	STRING	A JSON string that contains the predicted probability for each class. In the string, 0 represents the negative class and 1 represents the positive class. Example: {"0":0.1813110520,"1":0.8186889480}.