evalml.pipelines.TimeSeriesClassificationPipeline¶
-
class
evalml.pipelines.
TimeSeriesClassificationPipeline
(component_graph, parameters=None, custom_name=None, custom_hyperparameters=None, random_seed=0)[source]¶ Pipeline base class for time series classification problems.
Instance attributes
classes_
Gets the class names for the problem.
custom_hyperparameters
Custom hyperparameters for the pipeline.
custom_name
Custom name of the pipeline.
default_parameters
The default parameter dictionary for this pipeline.
feature_importance
Importance associated with each feature.
hyperparameters
Returns hyperparameter ranges from all components as a dictionary
linearized_component_graph
this is not guaranteed to be in proper component computation order
model_family
Returns model family of this pipeline template
name
Name of the pipeline.
parameters
Parameter dictionary for this pipeline
problem_type
summary
A short summary of the pipeline structure, describing the list of components used.
Methods:
Machine learning pipeline for time series classification problems made out of transformers and a classifier.
Determine whether the threshold of a binary classification pipeline can be tuned.
Constructs a new pipeline with the same components, parameters, and random state.
Transforms the data by applying all pre-processing components.
Outputs pipeline details including component parameters
Fit a time series classification pipeline.
Returns component by name
Generate an image representing the pipeline graph
Generate a bar graph of the pipeline’s feature importance
Loads pipeline at file path
Constructs a new instance of the pipeline with the same component graph but with a different set of parameters.
Make predictions using selected features.
Make probability estimates for labels.
Saves pipeline at file path
Evaluate model performance on current and additional objectives.