This function creates a custom feature from multiple numerical columns.
| Input Parameters | Type | Default | Description |
|---|
| cols | List[str] | | A list of column names that define this custom feature. |
| custom_name | str | | The name of this custom feature as it will appear in the monitoring tab. |
| transformation | str | None | An optional transformation step. Currently no transformation is available. |
| n_clusters | int | 10 | Number of clusters used for clustering-based drift monitoring. |
| monitor | bool | True | A 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 Type | Description |
|---|
| fdl.CustomFeature | A fdl.CustomFeature object which can be passed to model info. |
