Mli_autodoc configuration

make_autoreport

Make AutoDoc (Boolean) (Expert Setting)

Default value True

Whether to create the experiment AutoDoc after end of experiment.

autodoc_report_name

AutoDoc Name (String) (Expert Setting)

Default value 'report'

Specify the name of the report.

autodoc_template

AutoDoc Template Location (String) (Expert Setting)

Default value ''

AutoDoc template path. Provide the full path to your custom AutoDoc template or leave as ‘default’to generate the standard AutoDoc.

autodoc_additional_template_folder

autodoc_additional_template_folder (String)

Default value ''

Location of the additional AutoDoc templates

autodoc_output_type

AutoDoc File Output Type (String) (Expert Setting)

Default value 'docx'

Specify the AutoDoc output type.

autodoc_subtemplate_type

AutoDoc SubTemplate Type (String) (Expert Setting)

Default value 'auto'

Specify the type of sub-templates to use.

Options are ‘auto’, ‘docx’ or ‘md’.

autodoc_max_cm_size

Confusion Matrix Max Number of Classes (Number) (Expert Setting)

Default value 10

Specify the maximum number of classes in the confusion

matrix.

autodoc_num_features

Number of Top Features to Document (Number) (Expert Setting)

Default value 50

Specify the number of top features to display in

the document. setting to -1 disables this restriction.

autodoc_min_relative_importance

Minimum Relative Feature Importance Threshold (Float) (Expert Setting)

Default value 0.003

Specify the minimum relative importance in order

for a feature to be displayed. autodoc_min_relative_importance must be a float >= 0 and <= 1.

autodoc_include_permutation_feature_importance

Permutation Feature Importance (Boolean) (Expert Setting)

Default value False

Whether to compute permutation based feature

importance.

autodoc_feature_importance_num_perm

Number of Permutations for Feature Importance (Number) (Expert Setting)

Default value 1

Number of permutations to make per feature when computing

feature importance.

autodoc_feature_importance_scorer

Feature Importance Scorer (String) (Expert Setting)

Default value ''

Name of the scorer to be used to calculate feature

importance. Leave blank to use experiments default scorer.

autodoc_pd_max_rows

PDP and Shapley Summary Plot Max Rows (Number) (Expert Setting)

Default value 10000

The autodoc_pd_max_rows configuration controls the

number of rows shown for the partial dependence plots (PDP) and Shapley values summary plot in the AutoDoc. Random sampling is used for datasets with more than the autodoc_pd_max_rows limit.

autodoc_pd_max_runtime

PDP max runtime in seconds (Number) (Expert Setting)

Default value 45

Maximum number of seconds Partial Dependency computation

can take when generating report. Set to -1 for no time limit.

autodoc_pd_fast_approx

Use fast approximation for PDP (Boolean) (Expert Setting)

Default value True

Whether to enable fast approximation for predictions that are needed for the generation of partial dependence plots. Can help when want to create many PDP plots in short time. Amount of approximation is controlled by fast_approx_num_trees, fast_approx_do_one_fold, fast_approx_do_one_model experiment expert settings.

autodoc_pd_max_int_as_cat_uniques

PDP Max. number of unique values for int/float to be categoricals (Number) (Expert Setting)

Default value 50

Max number of unique values for integer/real columns to be treated as categoricals (test applies to first statistical_threshold_data_size_small rows only)

Similar to max_int_as_cat_uniques used for experiment, but here used to control PDP making.

autodoc_out_of_range

PDP Out of Range (Number) (Expert Setting)

Default value 3

Number of standard deviations outside of the range of

a column to include in partial dependence plots. This shows how the model will react to data it has not seen before.

autodoc_num_rows

ICE Number of Rows (Number) (Expert Setting)

Default value 0

Specify the number of rows to include in PDP and ICE plot

if individual rows are not specified.

autodoc_population_stability_index

Population Stability Index (Boolean) (Expert Setting)

Default value False

Whether to include population stability index if

experiment is binary classification/regression.

autodoc_population_stability_index_n_quantiles

Population Stability Index Number of Quantiles (Number) (Expert Setting)

Default value 10

Number of quantiles to use for population stability index

.

autodoc_prediction_stats

Prediction Statistics (Boolean) (Expert Setting)

Default value False

Whether to include prediction statistics information if

experiment is binary classification/regression.

autodoc_prediction_stats_n_quantiles

Prediction Statistics Number of Quantiles (Number) (Expert Setting)

Default value 20

Number of quantiles to use for prediction statistics.

autodoc_response_rate

Response Rates Plot (Boolean) (Expert Setting)

Default value False

Whether to include response rates information if

experiment is binary classification.

autodoc_response_rate_n_quantiles

Response Rate Plot Number of Quantiles (Number) (Expert Setting)

Default value 10

Number of quantiles to use for response rates information

.

autodoc_gini_plot

Show GINI Plot (Boolean) (Expert Setting)

Default value False

Whether to show the Gini Plot.

autodoc_enable_shapley_values

Enable Shapley Values (Boolean) (Expert Setting)

Default value True

Show Shapley values results in the AutoDoc.

autodoc_global_klime_num_features

Global KLIME Number of Features (Number) (Expert Setting)

Default value 10

The number feature in a KLIME global GLM coefficients

table. Must be an integer greater than 0 or -1. To show all features set to -1.

autodoc_global_klime_num_tables

Global KLIME Number of Tables (Number) (Expert Setting)

Default value 1

Set the number of KLIME global GLM coefficients tables. Set

to 1 to show one table with coefficients sorted by absolute value. Set to 2 to two tables one with the top positive coefficients and one with the top negative coefficients.

autodoc_data_summary_col_num

Number of Features in Data Summary Table (Number) (Expert Setting)

Default value -1

Number of features to be show in data summary. Value

must be an integer. Values lower than 1, f.e. 0 or -1, indicate that all columns should be shown.

autodoc_list_all_config_settings

List All Config Settings (Boolean) (Expert Setting)

Default value False

Whether to show all config settings. If False, only

the changed settings (config overrides) are listed, otherwise all settings are listed.

autodoc_keras_summary_line_length

Keras Model Architecture Summary Line Length (Number) (Expert Setting)

Default value -1

Line length of the keras model architecture summary. Must

be an integer greater than 0 or -1. To use the default line length set value -1.

autodoc_transformer_architecture_max_lines

NLP/Image Transformer Architecture Max Lines (Number) (Expert Setting)

Default value 30

Maximum number of lines shown for advanced transformer

architecture in the Feature section. Note that the full architecture can be found in the Appendix.

autodoc_full_architecture_in_appendix

Appendix NLP/Image Transformer Architecture (Boolean) (Expert Setting)

Default value False

Show full NLP/Image transformer architecture in

the Appendix.

autodoc_coef_table_appendix_results_table

Full GLM Coefficients Table in the Appendix (Boolean) (Expert Setting)

Default value False

Specify whether to show the full glm coefficient

table(s) in the appendix. coef_table_appendix_results_table must be a boolean: True to show tables in appendix, False to not show them .

autodoc_coef_table_num_models

GLM Coefficient Tables Number of Models (Number) (Expert Setting)

Default value 1

Set the number of models for which a glm coefficients

table is shown in the AutoDoc. coef_table_num_models must be -1 or an integer >= 1 (-1 shows all models).

autodoc_coef_table_num_folds

GLM Coefficient Tables Number of Folds Per Model (Number) (Expert Setting)

Default value -1

Set the number of folds per model for which a glm

coefficients table is shown in the AutoDoc. coef_table_num_folds must be -1 or an integer >= 1 (-1 shows all folds per model).

autodoc_coef_table_num_coef

**GLM Coefficient Tables Number of Coefficients ** (Number) (Expert Setting)

Default value 50

Set the number of coefficients to show within a glm

coefficients table in the AutoDoc. coef_table_num_coef, controls the number of rows shown in a glm table and must be -1 or an integer >= 1 (-1 shows all coefficients).

autodoc_coef_table_num_classes

GLM Coefficient Tables Number of Classes (Number) (Expert Setting)

Default value 9

Set the number of classes to show within a glm

coefficients table in the AutoDoc. coef_table_num_classes controls the number of class-columns shown in a glm table and must be -1 or an integer >= 4 (-1 shows all classes).

autodoc_num_histogram_plots

Number of Histograms to Show (Number) (Expert Setting)

Default value 10

When histogram plots are available: The number of

top (default 10) features for which to show histograms.

autodoc_force_singlenode

autodoc_force_singlenode (Boolean)

Default value False

If True, will force AutoDoc to run in only the main server, not on remote workers in case of a multi-node setup