{{ mlTaskDesign.taskType === 'PREDICTION' && isTimeseriesPrediction() ? 'External features' : 'Features Handling'}} Some of the selected algorithms do not use external features

# [ ]

Rejected, {{ featureAutoHandlingShortReason[feature.autoReason] }} {{ isCausalPrediction() | targetRoleName | capitalize }} variable Treatment variable Display only Dummy encoding None Target encoding Ordinal encoding Frequency encoding Presence flag Hashing Custom , drop missing , impute missing Presence flag Custom Avg-std rescaling Min-max rescaling Binarization Binning in {{feature.quantile_bin_nb_bins}} quantiles Cyclical datetime encoding Term hashing Term hashing + SVD TF/IDF vectorization Count vectorization Text embedding Custom Unfolding Custom Image embedding , drop missing , impute missing {{feature.sendToInput}}

No features matching filters

Handling of {{selection.selectedObjects.length}} selected features out of {{selection.allObjects.length}} available

Role

Type

Missing values handling

Numerical handling

Rescaling

Make derived features
(√x, x², log(x))

Category handling

Drop dummy

Text handling

Deep Learning Input Handling

{{ mlTaskFeatures(mlTaskDesign.preprocessing.per_feature, ['INPUT']).length || 'No' }} external {{ 'feature' | plurify: mlTaskFeatures(mlTaskDesign.preprocessing.per_feature, ['INPUT']).length }}

Some algorithms can use exogenous time-dependent factors (e.g. holidays, special events).

Future values of external features will be required to forecast.