Skip to content

MLI

MLIExplainerId

Bases: Enum

Enumeration represeting different MLI explainer IDs.

ABSOLUTE_PERMUTATION_BASED_FEATURE_IMPORTANCE class-attribute instance-attribute 

ABSOLUTE_PERMUTATION_BASED_FEATURE_IMPORTANCE = "h2oai.mli.byor.recipes.permutation_feat_imp_absolute_explainer.AbsolutePermutationFeatureImportanceExplainer"

Absolute Permutation-Based Feature Importance

DECISION_TREE class-attribute instance-attribute 

DECISION_TREE = "h2oaicore.mli.byor.recipes.surrogates.dt_surrogate_explainer.DecisionTreeSurrogateExplainer"

Decision Tree

FRIEDMAN_H_STATISTIC class-attribute instance-attribute 

FRIEDMAN_H_STATISTIC = "h2oai.mli.byor.recipes.h2o_sonar_explainers_friedman_h_statistic_explainer_FriedmanHStatisticExplainer"

Friedman's H-statistic

KERNEL_SHAPLEY_VALUES_FOR_ORIGINAL_FEATURES class-attribute instance-attribute 

KERNEL_SHAPLEY_VALUES_FOR_ORIGINAL_FEATURES = "h2oaicore.mli.byor.recipes.orig_kernel_shap_explainer.OriginalKernelShapExplainer"

Shapley Values for Original Features (Kernel SHAP Method)

NLP_LEAVE_ONE_COVARIATE_OUT class-attribute instance-attribute 

NLP_LEAVE_ONE_COVARIATE_OUT = "h2oaicore.mli.byor.recipes.text.nlp_loco_explainer_v2.NlpLocoExplainerVersion2"

NLP Leave-one-covariate-out (LOCO)

NLP_PARTIAL_DEPENDENCE_PLOT class-attribute instance-attribute 

NLP_PARTIAL_DEPENDENCE_PLOT = "h2oaicore.mli.byor.recipes.text.nlp_dai_pd_ice_explainer.NlpDaiPdIceExplainer"

NLP Partial Dependence Plot

NLP_TOKENIZER class-attribute instance-attribute 

NLP_TOKENIZER = "h2oaicore.mli.byor.recipes.text.nlp_tokenizer_explainer.NlpTokenizerExplainer"

NLP Tokenizer

NLP_VECTORIZER_LINEAR_MODEL class-attribute instance-attribute 

NLP_VECTORIZER_LINEAR_MODEL = "h2oaicore.mli.byor.recipes.text.nlp_vectorizer_linear_model_explainer.NlpVecLmExplainer"

NLP Vectorizer + Linear Model (VLM) Text Feature Importance

ORIGINAL_FEATURE_IMPORTANCE class-attribute instance-attribute 

ORIGINAL_FEATURE_IMPORTANCE = "h2oaicore.mli.byor.recipes.surrogates.original_feat_imp_explainer.OriginalFeatureImportanceExplainer"

Original Feature Importance

PARTIAL_DEPENDENCE_PLOT class-attribute instance-attribute 

PARTIAL_DEPENDENCE_PLOT = (
    "h2oaicore.mli.byor.recipes.dai_pd_ice_explainer.DaiPdIceExplainer"
)

Partial Dependence Plot

RANDOM_FOREST_FEATURE_IMPORTANCE class-attribute instance-attribute 

RANDOM_FOREST_FEATURE_IMPORTANCE = "h2oaicore.mli.byor.recipes.surrogates.rf_feat_imp_explainer.RandomForestFeatureImportanceExplainer"

Random Forest Feature Importance

RANDOM_FOREST_PARTIAL_DEPENDENCE_PLOT class-attribute instance-attribute 

RANDOM_FOREST_PARTIAL_DEPENDENCE_PLOT = "h2oaicore.mli.byor.recipes.surrogates.rf_pd_explainer.RandomForestPartialDependenceExplainer"

Random Forest Partial Dependence Plot

RELATIVE_PERMUTATION_BASED_FEATURE_IMPORTANCE class-attribute instance-attribute 

RELATIVE_PERMUTATION_BASED_FEATURE_IMPORTANCE = "h2oai.mli.byor.recipes.permutation_feat_imp_relative_explainer.RelativePermutationFeatureImportanceExplainer"

Relative Permutation-Based Feature Importance

SHAPLEY_SUMMARY_PLOT_FOR_ORIGINAL_FEATURES class-attribute instance-attribute 

SHAPLEY_SUMMARY_PLOT_FOR_ORIGINAL_FEATURES = "h2oaicore.mli.byor.recipes.shapley_summary_explainer.ShapleySummaryOrigFeatExplainer"

Shapley Summary Plot for Original Features (Naive Shapley Method)

SHAPLEY_VALUES_FOR_ORIGINAL_FEATURES class-attribute instance-attribute 

SHAPLEY_VALUES_FOR_ORIGINAL_FEATURES = "h2oaicore.mli.byor.recipes.original_contrib_explainer.NaiveShapleyExplainer"

Shapley Values for Original Features (Naive Method)

SHAPLEY_VALUES_FOR_TRANSFORMED_FEATURES class-attribute instance-attribute 

SHAPLEY_VALUES_FOR_TRANSFORMED_FEATURES = "h2oaicore.mli.byor.recipes.transformed_shapley_explainer.TransformedShapleyExplainer"

Shapley Values for Transformed Features

SURROGATE_RANDOM_FOREST_LEAVE_ONE_COVARIATE_OUT class-attribute instance-attribute 

SURROGATE_RANDOM_FOREST_LEAVE_ONE_COVARIATE_OUT = "h2oaicore.mli.byor.recipes.surrogates.rf_loco_explainer.RandomForestLocoExplainer"

Surrogate Random Forest Leave-one-covariate-out (LOCO)

TRANSFORMED_FEATURE_IMPORTANCE class-attribute instance-attribute 

TRANSFORMED_FEATURE_IMPORTANCE = "h2oaicore.mli.byor.recipes.transformed_feat_imp_explainer.TransformedFeatureImportanceExplainer"

Transformed Feature Importance

MLI

Interact with MLI interpretations on the Driverless AI server.

iid property 

iid: IIDMethods

Retrieve IID interpretations.

Deprecated API

A deprecated API that will be removed from v1.10.7 onwards. Please use 'driverlessai._mli.MLI' instead.

Returns:

  • IIDMethods –

     

create 

create(
    experiment: Optional[Experiment] = None,
    dataset: Optional[Dataset] = None,
    name: Optional[str] = None,
    force: bool = False,
    **kwargs: Any
) -> Interpretation

Create an MLI interpretation on the Driverless AI server and return an Interpretation object corresponding to the created interpretation.

Parameters:

  • experiment (Optional[Experiment], default: None ) –

    The experiment to interpret. Will use training dataset if the dataset has not been specified.

  • dataset (Optional[Dataset], default: None ) –

    The dataset to use for the interpretation (if dataset includes target and prediction columns, then an experiment will not be needed).

  • name (Optional[str], default: None ) –

    The display name for the interpretation.

  • force (bool, default: False ) –

    Create the new interpretation even if the interpretation with the same name already exists.

Other Parameters:

  • explainers (List[ExplainerRecipe]) –

    The list of explainer recipe objects.

  • test_dataset (Dataset) –

    Dataset object (timeseries only).

  • target_column (str) –

    The name of the column in dataset.

  • prediction_column (str) –

    The name of the column in dataset.

  • weight_column (str) –

    The name of the column in dataset.

  • drop_columns (List[str]) –

    The names of the columns in dataset.

Note

Any expert setting can also be passed as a kwarg. To search possible expert settings for your server version, use mli.search_expert_settings(search_term).

Returns:

create_async 

create_async(
    experiment: Optional[Experiment] = None,
    dataset: Optional[Dataset] = None,
    name: Optional[str] = None,
    force: bool = False,
    validate_settings: bool = True,
    **kwargs: Any
) -> Interpretation

Launch creation of an MLI interpretation on the Driverless AI server and return an Interpretation object to track the status.

Parameters:

  • experiment (Optional[Experiment], default: None ) –

    The experiment to interpret. Will use training dataset if the dataset has not been specified.

  • dataset (Optional[Dataset], default: None ) –

    The dataset to use for the interpretation (if dataset includes target and prediction columns, then an experiment will not be needed).

  • name (Optional[str], default: None ) –

    The display name for the interpretation.

  • force (bool, default: False ) –

    Create the new interpretation even if the interpretation with the same name already exists.

Other Parameters:

  • explainers (List[ExplainerRecipe]) –

    The list of explainer recipe objects.

  • test_dataset (Dataset) –

    Dataset object (timeseries only).

  • target_column (str) –

    The name of the column in dataset.

  • prediction_column (str) –

    The name of the column in dataset.

  • weight_column (str) –

    The name of the column in dataset.

  • drop_columns (List[str]) –

    The names of the columns in dataset.

Note

Any expert setting can also be passed as a kwarg. To search possible expert settings for your server version, use mli.search_expert_settings(search_term).

Returns:

get 

get(key: str) -> Interpretation

Initialize an Interpretation object but do not request information from the server (it is possible for the interpretation key to not exist on the server). This is useful for populating lists without making multiple network calls.

Parameters:

  • key (str) –

    Driverless AI server's unique ID for the MLI interpretation.

Returns:

gui 

gui() -> Hyperlink

Print the full URL for the user's MLI page on Driverless AI server.

Returns:

list 

list(start_index: int = 0, count: int = None) -> Sequence[Interpretation]

List of Interpretation objects available to the user.

Parameters:

  • start_index (int, default: 0 ) –

    The index number on Driverless AI server of the first interpretation to in the list.

  • count (int, default: None ) –

    The max number of interpretations to request from the Driverless AI server.

Returns:

search_expert_settings 

search_expert_settings(
    search_term: str = "",
    show_description: bool = False,
    show_valid_values: bool = False,
) -> Table

Search expert settings and print results. Useful when looking for kwargs to use when creating interpretations.

Parameters:

  • search_term (str, default: '' ) –

    Term to search for (case-insensitive).

  • show_description (bool, default: False ) –

    Include description in results.

  • show_valid_values (bool, default: False ) –

    Include the valid values that can be set for each setting in the results.

Returns:

Interpretation

Interact with an MLI interpretation on the Driverless AI server.

artifacts property 

artifacts: InterpretationArtifacts

Interact with artifacts that are created when the interpretation completes.

Returns:

creation_timestamp property 

creation_timestamp: float

Creation timestamp in seconds since the epoch (POSIX timestamp).

Returns:

dataset property 

dataset: Optional[Dataset]

Dataset for the interpretation.

Returns:

experiment property 

experiment: Optional[Experiment]

Experiment for the interpretation.

Returns:

explainers property 

explainers: ExplainerList

Explainers that were ran as an ExplainerList object.

Driverless AI version requirement

Requires Driverless AI server 1.10.5 or higher.

Beta API

A beta API that is subject to future changes.

Returns:

explanation_plots property 

explanation_plots: Mapping[Union[int, str], List[ExplanationPlot]]

Plots for explanations that were created for the interpretation.

Example: Retrieve global and local explanation plot 
# get the list of available plots for the Decision Tree explainer
MLIExplainerId = driverlessai.MLIExplainerId
dt_plots = interpretation.explanation_plots[MLIExplainerId.DECISION_TREE]

# retrieve the global explanation
dt_plot = dt_plots[0].get_plot()

# get the local explanation (for row 44)
dt_local_plot = dt_plots[0].get_plot(row_number=44)
Driverless AI version requirement

Requires Driverless AI server 1.10.5 or higher.

Beta API

A beta API that is subject to future changes.

Returns:

key property 

key: str

Universally unique key of the entity.

Returns:

name property 

name: str

Name of the entity.

Returns:

run_duration property 

run_duration: Optional[float]

Run duration in seconds.

Returns:

settings property 

settings: Dict[str, Any]

Interpretation settings.

Returns:

delete 

delete() -> None

Delete MLI interpretation on Driverless AI server.

gui 

gui() -> Hyperlink

Get the full URL for the interpretation's page on the Driverless AI server.

Returns:

is_complete 

is_complete() -> bool

Whether the job has been completed successfully.

Returns:

  • bool –

    True if the job has been completed successfully, otherwise False.

is_running 

is_running() -> bool

Whether the job has been scheduled or is running, finishing, or syncing.

Returns:

  • bool –

    True if the job has not completed yet, otherwise False.

parameter_summary 

parameter_summary() -> Table

Get MLI summary.

Driverless AI version requirement

Requires Driverless AI server 1.10.5 or higher.

Returns:

rename 

rename(name: str) -> Interpretation

Change the interpretation display name.

Parameters:

  • name (str) –

    New display name.

Returns:

result 

result(silent: bool = False) -> Interpretation

Wait for the job to complete, then return an Interpretation object.

Returns:

status 

status(verbose: int = 0) -> str

Returns the status of the job.

Parameters:

  • verbose (int, default: 0 ) –
    • 0: A short description.
    • 1: A short description with a progress percentage.
    • 2: A detailed description with a progress percentage.

Returns:

  • str –

    Current status of the job.

InterpretationArtifacts

Interact with files created by an MLI interpretation on the Driverless AI server.

file_paths property 

file_paths: Dict[str, str]

Paths to interpretation artifact files on the server.

Returns:

download 

download(
    only: Union[str, List[str]] = None,
    dst_dir: str = ".",
    file_system: Optional[AbstractFileSystem] = None,
    overwrite: bool = False,
) -> Dict[str, str]

Download interpretation artifacts from the Driverless AI server. Returns a dictionary of relative paths for the downloaded artifacts.

Parameters:

  • only (Union[str, List[str]], default: None ) –

    Specify the specific artifacts to download, use interpretation.artifacts.list() to see the available artifacts on the Driverless AI server.

  • dst_dir (str, default: '.' ) –

    The path to the directory where the interpretation artifacts will be saved.

  • file_system (Optional[AbstractFileSystem], default: None ) –

    FSSPEC based file system to download to, instead of local file system.

  • overwrite (bool, default: False ) –

    Overwrite existing files.

Returns:

list 

list() -> List[str]

List of interpretation artifacts that exist on the Driverless AI server.

Returns:

Explainer

Interact with an MLI explainers on the Driverless AI server.

artifacts property 

artifacts: ExplainerArtifacts

Artifacts of this explainer.

Returns:

explanation_plots property 

explanation_plots: List[ExplanationPlot]

The avaialbe plots for this explainer.

Returns:

frames property 

frames: Optional[ExplainerFrames]

An ExplainerFrames object that contains the paths to the explainer frames retrieved from Driverless AI Server. If the explainer frame is not available, the value of this propertiy is None.

Returns:

id property 

id: str

This explainer's Id.

Returns:

key property 

key: str

Universally unique key of the entity.

Returns:

name property 

name: str

Name of the entity.

Returns:

get_data 

get_data(**kwargs: Any) -> ExplainerData

Retrieve the ExplainerData from the Driverless AI server. Raises a RuntimeError exception if the explainer has not been completed successfully.

Use help(explainer.get_data) to view help on available keyword arguments.

Returns:

is_complete 

is_complete() -> bool

Whether the job has been completed successfully.

Returns:

  • bool –

    True if the job has been completed successfully, otherwise False.

is_running 

is_running() -> bool

Whether the job has been scheduled or is running, finishing, or syncing.

Returns:

  • bool –

    True if the job has not completed yet, otherwise False.

result 

result(silent: bool = False) -> Explainer

Wait for the explainer to complete, then return self.

Parameters:

  • silent (bool, default: False ) –

    If True, do not display status updates.

Returns:

status 

status(verbose: int = 0) -> str

Returns the status of the job.

Parameters:

  • verbose (int, default: 0 ) –
    • 0: A short description.
    • 1: A short description with a progress percentage.
    • 2: A detailed description with a progress percentage.

Returns:

  • str –

    Current status of the job.

ExplainerArtifacts

Interact with artifacts created by an explainer during interpretation on the Driverless AI server.

file_paths property 

file_paths: Dict[str, str]

Paths to explainer artifact files on the server.

Returns:

download 

download(
    only: Union[str, List[str]] = None,
    dst_dir: str = ".",
    file_system: Optional[AbstractFileSystem] = None,
    overwrite: bool = False,
) -> Dict[str, str]

Download explainer artifacts from the Driverless AI server. Returns a dictionary of relative paths for the downloaded artifacts.

Parameters:

  • only (Union[str, List[str]], default: None ) –

    Specify the specific artifacts to download, use interpretation.artifacts.list() to see the available artifacts on the Driverless AI server.

  • dst_dir (str, default: '.' ) –

    The path to the directory where the interpretation artifacts will be saved.

  • file_system (Optional[AbstractFileSystem], default: None ) –

    FSSPEC based file system to download to, instead of local file system.

  • overwrite (bool, default: False ) –

    Overwrite existing files.

Returns:

list 

list() -> List[str]

List of explainer artifacts that exist on the Driverless AI server.

Returns:

ExplainerData

Interact with the result data of an explainer on the Driverless AI server.

data property 

data: str

The explainer result data as string.

Returns:

data_format property 

data_format: str

The explainer data format.

Returns:

data_type property 

data_type: str

The explainer data type.

Returns:

data_as_dict 

data_as_dict() -> Optional[Union[List, Dict]]

Return the explainer result data as a dictionary.

Returns:

data_as_pandas 

data_as_pandas() -> Optional[DataFrame]

Return the explainer result data as a pandas frame.

Beta API

A beta API that is subject to future changes.

Returns:

ExplainerFrames

Interact with explanation frames created by an explainer during interpretation on the Driverless AI server.

frame_paths property 

frame_paths: Dict[str, str]

Frame names and paths to artifact files on the server.

Returns:

download 

download(
    frame_name: Union[str, List[str]] = None,
    dst_dir: str = ".",
    file_system: Optional[AbstractFileSystem] = None,
    overwrite: bool = False,
) -> Dict[str, str]

Download explainer frames from the Driverless AI server. Returns a dictionary of relative paths for the downloaded artifacts.

Parameters:

  • frame_name (Union[str, List[str]], default: None ) –

    Specify the specific frame to download, use explainer.frames.list() to see the available artifacts on the Driverless AI server.

  • dst_dir (str, default: '.' ) –

    The path to the directory where the interpretation artifacts will be saved.

  • file_system (Optional[AbstractFileSystem], default: None ) –

    Optional["fsspec.spec.AbstractFileSystem"] = None.

  • overwrite (bool, default: False ) –

    Overwrite existing files.

Returns:

frame_as_pandas 

frame_as_pandas(
    frame_name: str,
    custom_tmp_dir: Optional[str] = None,
    keep_downloaded: bool = False,
) -> DataFrame

Download a frame with the given frame name to a temporary directory and return it as a pandas.DataFrame.

Parameters:

  • frame_name (str) –

    The name of the frame to open.

  • custom_tmp_dir (Optional[str], default: None ) –

    If specified, use this directory as the temporary directory instead of the default.

  • keep_downloaded (bool, default: False ) –

    If True, do not delete the downloaded frame. Otherwise, the downloaded frame is deleted before returning from this method.
Beta API

A beta API that is subject to future changes.

Returns:

frame_names 

frame_names() -> List[str]

List of explainer frames that exist on the Driverless AI server.

Returns:

ExplainerList

List that lazy loads Explainer objects.

__get_by_index 

__get_by_index(idx: int) -> Explainer

Finds the explainer object that corresponds to the given index, and initializes it if it is not already initialized.

Parameters:

  • index –

    The index of the desired explainer.

get_by_key 

get_by_key(key: str) -> Explainer

Finds the explainer object that corresponds to the given key, and initializes it if it is not already initialized.

Parameters:

  • key (str) –

    The job key of the desired explainer.
Beta API

A beta API that is subject to future changes.

Returns:

get_by_name 

get_by_name(name: str) -> Explainer

Finds the explainer object that corresponds to the given explainer name, and initializes it if it is not already initialized.

Parameters:

  • key –

    The name of the desired explainer.
Beta API

A beta API that is subject to future changes.

Returns:

ExplanationPlot

This abstract class serves as the foundation for all explanation plot classes to inherit from.

ShapleyExplanationPlot

Handles plots for the following explainers:

  • Shapley Values for Original Features (Naive Method)
  • Shapley Values for Transformed Features
  • Shapley Values for Original Features (Kernel SHAP Method)

get_plot 

get_plot(
    *, class_name: Optional[str] = None, row_number: Optional[int] = None
) -> Dict[str, Any]

Plot this explanation

Parameters:

  • row_number (Optional[int], default: None ) –

    Local explanation for the given row_number.

  • class_name (Optional[str], default: None ) –

    The name of the class in multinomial classification, if not provided the first class from the set of available classes in the model will be selected (use the method get_classes() to view available classes).

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format

NLPTokenizerExplanationPlot

Handles plots for the following explainers:

  • NLP Tokenizer

get_plot 

get_plot(*, row_number: Optional[int] = None) -> Dict[str, Any]

Plot this explanation

Parameters:

  • row_number (Optional[int], default: None ) –

    Local explanation for the given row_number.

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format

FeatureImportanceGlobalExplanationPlot

Handles plots for the following explainers:

  • Absolute Permutation-Based Feature Importance
  • Relative Permutation-Based Feature Importance
  • Friedman's H-statistic

get_plot 

get_plot(*, class_name: Optional[str] = None) -> Dict[str, Any]

Plot this explanation

Parameters:

  • class_name (Optional[str], default: None ) –

    The name of the class in multinomial classification, if not provided the first class from the set of available classes in the model will be selected (use the method get_classes() to view available classes).

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format

DAIFeatureImportanceExplanationPlot

Handles plots for the following explainers:

  • Original Feature Importance
  • Transformed Feature Importance

get_plot 

get_plot() -> Dict[str, Any]

Plot this explanation

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format

RandomForestFeatureImportanceExplanationPlot

Handles plots for the following explainers:

  • Random Forest Feature Importance
  • Surrogate Random Forest Leave-one-covariate-out (LOCO)

get_plot 

get_plot(*, row_number: Optional[int] = None) -> Dict[str, Any]

Plot this explanation

Parameters:

  • row_number (Optional[int], default: None ) –

    Local explanation for the given row_number.

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format

NLPFeatureImportanceExplanationPlot

Handles plots for the following explainers:

  • NLP Vectorizer + Linear Model (VLM) Text Feature Importance
  • NLP Leave-one-covariate-out (LOCO)

get_plot 

get_plot(
    *,
    row_number: Optional[int] = None,
    class_name: Optional[str] = None,
    text_feature: Optional[str] = None
) -> Dict[str, Any]

Plot this explanation

Parameters:

  • row_number (Optional[int], default: None ) –

    Local explanation for the given row_number.

  • class_name (Optional[str], default: None ) –

    The name of the class in multinomial classification, if not provided the first class from the set of available classes in the model will be selected (use the method get_classes() to view available classes).

  • text_features –

    Select the text feature to plot, if not provided all text features will be selected (use the method get_text_features to view available text features)

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format

PDPExplanationPlot

Handles plots for the following explainers:

  • Partial Dependence Plot
  • NLP Partial Dependence Plot

get_plot 

get_plot(
    *,
    row_number: Optional[int] = None,
    class_name: Optional[str] = None,
    feature_name: Optional[str] = None,
    partial_dependence_type: Optional[str] = None
) -> Dict[str, Any]

Plot this explanation

Parameters:

  • row_number (Optional[int], default: None ) –

    Local explanation for the given row_number.

  • class_name (Optional[str], default: None ) –

    The name of the class in multinomial classification, if not provided the first class from the set of available classes in the model will be selected (use the method get_classes() to view available classes).

  • feature_name (Optional[str], default: None ) –

    The name of the feature to plot, if not provided the first feature from the set of avaialbe features in the model will be selected (use the method get_feature_names() to view available feature names)

  • partial_dependence_type (Optional[str], default: None ) –

    Override default plot type, available options are categorical and numeric

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format

RandomForestPDPExplanationPlot

Handles plots for the following explainer:

  • Random Forest Partial Dependence Plot

get_plot 

get_plot(
    *, row_number: Optional[int] = None, feature_name: Optional[str] = None
) -> Dict[str, Any]

Plot this explanation

Parameters:

  • row_number (Optional[int], default: None ) –

    Local explanation for the given row_number.

  • feature_name (Optional[str], default: None ) –

    The name of the feature to plot, if not provided the first feature from the set of avaialbe features in the model will be selected (use the method get_feature_names() to view available feature names)

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format

DecisionTreeExplanationPlot

Handles plots for the following explainer:

  • Decision Tree

get_plot 

get_plot(
    *, row_number: Optional[int] = None, class_name: Optional[str] = None
) -> Dict[str, Any]

Plot this explanation

Parameters:

  • row_number (Optional[int], default: None ) –

    Local explanation for the given row_number.

  • class_name (Optional[str], default: None ) –

    The name of the class in multinomial classification, if not provided the first class from the set of available classes in the model will be selected (use the method get_classes() to view available classes).

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format

ShapleySummaryExplanationPlot

Handles plots for the following explainer:

  • Shapley Summary Plot for Original Features

get_plot 

get_plot(
    *, class_name: Optional[str] = None, page_num: int = 1
) -> Dict[str, Any]

Plot this explanation

Parameters:

  • class_name (Optional[str], default: None ) –

    The name of the class in multinomial classification, if not provided the first class from the set of available classes in the model will be selected (use the method get_classes() to view available classes).

  • page_num (int, default: 1 ) –

    Select the page (use the method get_total_pages() to view the total available number of pages)

Returns:

  • Dict[str, Any] –

    The plot in Vega Lite (v5) format