Explainer (레시피) 상세 설정¶

다음은 새로운 해석을 설정할 때 사용할 수 있는 Explainer별 상세 설정의 목록입니다. 이러한 설정은 recipes 탭 아래 MLI page 에서 해석을 실행할 때 액세스할 수 있습니다. 일반적인 MLI 상세 설정에 대한 정보는 해석 상세 설정 를 참조하십시오.

Absolute Permutation Feature Importance Explainer Settings
MLI AutoDoc Explainer Settings
Disparate Impact Analysis Explainer 설정
NLP Partial Dependence Plot Explainer Settings
NLP 벡터라이저 + 선형 모델 텍스트 기능 중요성 설명자 설정
partial dependence plot Explainer 설정
Sensitivity Analysis Explainer Settings
Shapley Summary Plot Explainer 설정
원래 기능 설정을 위한 Shapley Values
Surrogate Decision Tree Explainer 설정

Interpretation Explainer Expert Settings

Absolute Permutation Feature Importance Explainer Settings¶

`mli_sample_size`¶

`missing_values`¶

`autodoc_feature_importance_num_perm`¶

`autodoc_feature_importance_scorer`¶

MLI AutoDoc Explainer Settings¶

`autodoc_report_name`¶

`autodoc_template`¶

`autodoc_output_type`¶

`autodoc_subtemplate_type`¶

`autodoc_max_cm_size`¶

`autodoc_num_features`¶

`autodoc_min_relative_importance`¶

`autodoc_include_permutation_feature_importance`¶

`autodoc_feature_importance_num_perm`¶

`autodoc_feature_importance_scorer`¶

`autodoc_pd_max_rows`¶

`autodoc_pd_max_runtime`¶

`autodoc_out_of_range`¶

`autodoc_num_rows`¶

`autodoc_population_stability_index`¶

`autodoc_population_stability_index_n_quantiles`¶

`autodoc_prediction_stats`¶

`autodoc_prediction_stats_n_quantiles`¶

`autodoc_response_rate`¶

`autodoc_response_rate_n_quantiles`¶

`autodoc_gini_plot`¶

`autodoc_enable_shapley_values`¶

`autodoc_global_klime_num_features`¶

`autodoc_global_klime_num_tables`¶

`autodoc_data_summary_col_num`¶

`autodoc_list_all_config_settings`¶

`autodoc_keras_summary_line_length`¶

`autodoc_transformer_architecture_max_lines`¶

`autodoc_full_architecture_in_appendix`¶

`autodoc_coef_table_appendix_results_table`¶

`autodoc_coef_table_num_models`¶

`autodoc_coef_table_num_folds`¶

`autodoc_coef_table_num_coef`¶

`autodoc_coef_table_num_classes`¶

`autodoc_num_histogram_plots`¶

Disparate Impact Analysis Explainer 설정¶

Driverless AI의 Disparate Impact Analysis에 대한 정보는 Disparate Impact Analysis(DIA) 를 참조하십시오. 다음은 새로운 해석을 실행할 때 MLI 페이지의 레시피 탭에서 토글할 수 있는 매개변수의 목록입니다.

`dia_cols`¶

`cut_off`¶

`maximize_metric`¶

`use_holdout_preds`¶

`sample_size`¶

`max_card`¶

`min_card`¶

`num_card`¶

`fast_approx`¶

NLP Partial Dependence Plot Explainer Settings¶

`max_tokens`¶

`custom_tokens`¶

NLP 벡터라이저 + 선형 모델 텍스트 기능 중요성 설명자 설정¶

`txt_cols`¶

`cut_off`¶

`maximize_metric`¶

partial dependence plot Explainer 설정¶

Driverless AI의 부분 의존도 플롯에 대한 정보는 부분 의존도 플롯(PDP) 를 참조하십시오. 다음은 새로운 해석을 실행할 때 MLI 페이지의 레시피 탭에서 토글할 수 있는 매개변수의 목록입니다.

`sample_size`¶

`max_features`¶

`features`¶

`oor_grid_resolution`¶

`qtile_grid_resolution`¶

`grid_resolution`¶

`center`¶

`sort_bins`¶

`histograms`¶

`qtile-bins`¶

`1_frame`¶

`numcat_num_chart`¶

`numcat_threshold`¶

Sensitivity Analysis Explainer Settings¶

`sample_size`¶

Shapley Summary Plot Explainer 설정¶

Driverless AI의 Shapley Summary 플롯에 대한 자세한 내용은 Shapley Summary Plot (Original Features) 를 참조하십시오. 다음은 새로운 해석을 실행할 때 MLI 페이지의 레시피 탭에서 토글할 수 있는 매개변수의 목록입니다.

`max_features`¶

`sample_size`¶

`x_resolution`¶

`drilldown_charts`¶

`fast_approx`¶

원래 기능 설정을 위한 Shapley Values¶

`sample_size`¶

행의 수가 이 제한값을 초과하면 Naive Shapley에 대해 샘플링을 수행합니다. 기본값은 100000으르 설정됩니다.

`fast_approx`¶

Surrogate Decision Tree Explainer 설정¶

Driverless AI의 Surrogate Decision Tree Plot에 대한 자세한 내용은 Surrogate Decision Tree 를 참조하십시오. 다음은 새로운 해석을 실행할 때 MLI 페이지의 레시피 탭에서 토글할 수 있는 매개변수의 목록입니다.

`dt_tree_depth`¶

`nfolds`¶

`qbin_cols`¶

`qbin_count`¶

Explainer (레시피) 상세 설정¶

Absolute Permutation Feature Importance Explainer Settings¶

mli_sample_size¶

missing_values¶

autodoc_feature_importance_num_perm¶

autodoc_feature_importance_scorer¶

MLI AutoDoc Explainer Settings¶

autodoc_report_name¶

autodoc_template¶

autodoc_output_type¶

autodoc_subtemplate_type¶

autodoc_max_cm_size¶

autodoc_num_features¶

autodoc_min_relative_importance¶

autodoc_include_permutation_feature_importance¶

autodoc_feature_importance_num_perm¶

autodoc_feature_importance_scorer¶

autodoc_pd_max_rows¶

autodoc_pd_max_runtime¶

autodoc_out_of_range¶

autodoc_num_rows¶

autodoc_population_stability_index¶

autodoc_population_stability_index_n_quantiles¶

autodoc_prediction_stats¶

autodoc_prediction_stats_n_quantiles¶

autodoc_response_rate¶

autodoc_response_rate_n_quantiles¶

autodoc_gini_plot¶

autodoc_enable_shapley_values¶

autodoc_global_klime_num_features¶

autodoc_global_klime_num_tables¶

autodoc_data_summary_col_num¶

autodoc_list_all_config_settings¶

autodoc_keras_summary_line_length¶

autodoc_transformer_architecture_max_lines¶

autodoc_full_architecture_in_appendix¶

autodoc_coef_table_appendix_results_table¶

autodoc_coef_table_num_models¶

autodoc_coef_table_num_folds¶

autodoc_coef_table_num_coef¶

autodoc_coef_table_num_classes¶

autodoc_num_histogram_plots¶

Disparate Impact Analysis Explainer 설정¶

dia_cols¶

cut_off¶

maximize_metric¶

use_holdout_preds¶

sample_size¶

max_card¶

min_card¶

num_card¶

fast_approx¶

NLP Partial Dependence Plot Explainer Settings¶

max_tokens¶

custom_tokens¶

NLP 벡터라이저 + 선형 모델 텍스트 기능 중요성 설명자 설정¶

txt_cols¶

cut_off¶

maximize_metric¶

partial dependence plot Explainer 설정¶

sample_size¶

max_features¶

features¶

oor_grid_resolution¶

qtile_grid_resolution¶

grid_resolution¶

center¶

sort_bins¶

histograms¶

qtile-bins¶

1_frame¶

numcat_num_chart¶

numcat_threshold¶

Sensitivity Analysis Explainer Settings¶

sample_size¶

Shapley Summary Plot Explainer 설정¶

max_features¶

sample_size¶

x_resolution¶

drilldown_charts¶

`mli_sample_size`¶

`missing_values`¶

`autodoc_feature_importance_num_perm`¶

`autodoc_feature_importance_scorer`¶

`autodoc_report_name`¶

`autodoc_template`¶

`autodoc_output_type`¶

`autodoc_subtemplate_type`¶

`autodoc_max_cm_size`¶

`autodoc_num_features`¶

`autodoc_min_relative_importance`¶

`autodoc_include_permutation_feature_importance`¶

`autodoc_feature_importance_num_perm`¶

`autodoc_feature_importance_scorer`¶

`autodoc_pd_max_rows`¶

`autodoc_pd_max_runtime`¶

`autodoc_out_of_range`¶

`autodoc_num_rows`¶

`autodoc_population_stability_index`¶

`autodoc_population_stability_index_n_quantiles`¶

`autodoc_prediction_stats`¶

`autodoc_prediction_stats_n_quantiles`¶

`autodoc_response_rate`¶

`autodoc_response_rate_n_quantiles`¶

`autodoc_gini_plot`¶

`autodoc_enable_shapley_values`¶

`autodoc_global_klime_num_features`¶

`autodoc_global_klime_num_tables`¶

`autodoc_data_summary_col_num`¶

`autodoc_list_all_config_settings`¶

`autodoc_keras_summary_line_length`¶

`autodoc_transformer_architecture_max_lines`¶

`autodoc_full_architecture_in_appendix`¶

`autodoc_coef_table_appendix_results_table`¶

`autodoc_coef_table_num_models`¶

`autodoc_coef_table_num_folds`¶

`autodoc_coef_table_num_coef`¶

`autodoc_coef_table_num_classes`¶

`autodoc_num_histogram_plots`¶

`dia_cols`¶

`cut_off`¶

`maximize_metric`¶

`use_holdout_preds`¶

`sample_size`¶

`max_card`¶

`min_card`¶

`num_card`¶

`fast_approx`¶

`max_tokens`¶

`custom_tokens`¶

`txt_cols`¶

`cut_off`¶

`maximize_metric`¶

`sample_size`¶

`max_features`¶

`features`¶

`oor_grid_resolution`¶

`qtile_grid_resolution`¶

`grid_resolution`¶

`center`¶

`sort_bins`¶

`histograms`¶

`qtile-bins`¶

`1_frame`¶

`numcat_num_chart`¶

`numcat_threshold`¶

`sample_size`¶

`max_features`¶

`sample_size`¶

`x_resolution`¶

`drilldown_charts`¶

`fast_approx`¶

`sample_size`¶

`fast_approx`¶

`dt_tree_depth`¶

`nfolds`¶

`qbin_cols`¶

`qbin_count`¶