This function creates a custom feature from multiple numerical columns.

Input ParametersTypeDefaultDescription
colsList[str]A list of column names that define this custom feature.
custom_namestrThe name of this custom feature as it will appear in the monitoring tab.
transformationstrNoneAn optional transformation step. Currently no transformation is available.
n_clustersint10Number of clusters used for clustering-based drift monitoring.
monitorboolTrueA boolean variable that specifies whether this custom feature will be monitored using the clustering-based binning.
CF1 = fdl.CustomFeature.from_columns(['f1','f2','f3'], 
                                     custom_name = 'vector1',
                                     n_clusters=5)

model_info = fdl.ModelInfo.from_dataset_info(
    dataset_info=dataset_info,
    dataset_id = DATASET_ID,
    features = data_cols,
    target='target',
    outputs='predicted_score',
    custom_features = [CF1]
)
Return TypeDescription
fdl.CustomFeatureA fdl.CustomFeature object which can be passed to model info.