🕷️ Crawler Inspector

URL Lookup

Direct Parameter Lookup

Raw Queries and Responses

1. Shard Calculation

Query:

Response:

Calculated Shard: 169 (from laksa187)

2. Crawled Status Check

Query:

curl -X POST \
  'http://laksa169.int.ahrefs:8124/' \
  -H 'Content-Type: text/plain' \
  -H 'X-ClickHouse-Database: crawler3' \
  -H 'Authorization: Basic YXBpOg==' \
  -d 'SELECT getAhrefsURLFromUnparsed(src_unparsed) AS found_url, ifNull(toUnixTimestamp(download_stamp), 0) AS crawl_time, ifNull(toUnixTimestamp(props_url_first_seen), 0) AS first_indexed_time, download_http_code AS http_code, src_unparsed AS src_unparsed, src_root_hash AS src_root_hash, history_drop_reason AS history_drop_reason, meta_title AS meta_title, meta_descriptions AS meta_descriptions, attrs_boilerpipe_text AS attrs_boilerpipe_text, attrs_markdown AS attrs_markdown, attrs_readable_markdown AS attrs_readable_markdown, meta_canonical AS meta_canonical FROM crawler3.page_info_local FINAL PREWHERE (src_root_hash, src_unparsed) IN ((getAhrefsRootHashFromUnparsed(getAhrefsUnparsedNoserviceFromURL(\'https://catboost.ai/docs/en/concepts/python-reference_catboostclassifier_get_feature_importance\')), getAhrefsUnparsedNoserviceFromURL(\'https://catboost.ai/docs/en/concepts/python-reference_catboostclassifier_get_feature_importance\'))) FORMAT JSONEachRow'

Response:

{"found_url":"https:\/\/catboost.ai\/docs\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance","crawl_time":1775285507,"first_indexed_time":1731946335,"http_code":200,"src_unparsed":"ai,catboost!\/docs\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance s443","src_root_hash":"17435841955170310369","history_drop_reason":null,"meta_title":"get_feature_importance | CatBoost","meta_descriptions":["Calculate and return the feature importances. Method call format."],"attrs_boilerpipe_text":"Calculate and return the \nfeature importances\n.\nMethod call format\nget_feature_importance(data=\nNone\n,\n                       reference_data=\nNone\n,\ntype\n=EFstrType.FeatureImportance,\n                       prettified=\nFalse\n,\n                       thread_count=-\n1\n,\n                       verbose=\nFalse\n,\n                       log_cout=sys.stdout,\n                       log_cerr=sys.stderr)\nParameters\ndata\nDescription\nThe dataset for feature importance calculation.\nThe required dataset depends on the selected feature importance calculation type (specified in the\ntype\nparameter):\nPredictionValuesChange —  Either None or the same dataset that was used for training if the model does not contain information regarding the weight of leaves. All models trained with CatBoost version 0.9 or higher contain leaf weight information by default.\nLossFunctionChange  —  Any dataset. Feature importances are calculated on a subset for large datasets.\nPredictionDiff —  A list of object pairs.\nPossible types\ncatboost.Pool\nDefault value\nRequired parameter for the LossFunctionChange and ShapValues type of feature importances and in case the model does not contain information regarding the weight of leaves.\nNone otherwise.\nreference_data\nDescription\nReference data for Independent Tree SHAP values from\nExplainable AI for Trees: From Local Explanations to Global Understanding\n. If\ntype\nis\nShapValues\nand\nreference_data\nis not\nNone\n, then Independent Tree SHAP values are calculated.\nPossible types\ncatboost.Pool\nDefault value\nNone\ntype\nAlias:\nfstr_type\n(deprecated, use\ntype\ninstead)\nDescription\nThe type of feature importance to calculate.\nPossible values:\nFeatureImportance: Equal to\nPredictionValuesChange\nfor non-ranking metrics and\nLossFunctionChange\nfor ranking metrics (the value is determined automatically).\nShapValues\n: A vector\nv\nv\nwith contributions of each feature to the prediction for every input object and the expected value of the model prediction for the object (average prediction given no knowledge about the object).\nInteraction\n: The value of the feature interaction strength for each pair of features.\nPredictionDiff: A vector with contributions of each feature to the RawFormulaVal difference for each pair of objects.\nPossible types\nstring\nEFStrType\nNote\nIt is recommended to use EFStrType for this parameter.\nDefault value\nFeatureImportance\nprettified\nDescription\nReturn the feature importances as a list of the following pairs sorted by feature importance:\n(feature_id, feature importance)\nShould be used if one of the following values of the\ntype\nparameter is selected:\nPredictionValuesChange\nPredictionValuesChange\nPossible types\nbool\nDefault value\nFalse\nthread_count\nDescription\nThe number of threads to use for operation.\nOptimizes the speed of execution. This parameter doesn't affect results.\nPossible types\nint\nDefault value\n-1 (the number of threads is equal to the number of processor cores)\nverbose\nDescription\nThe purpose of this parameter depends on the type of the given value:\nbool —  Output progress to stdout.\nWorks with the\nShapValues\ntype of feature importance calculation.\nint —  The logging period.\nPossible types\nbool\nint\nDefault value\nFalse\nlog_cout\nOutput stream or callback for logging.\nPossible types\ncallable Python object\npython object providing the\nwrite()\nmethod\nDefault value\nsys.stdout\nlog_cerr\nError stream or callback for logging.\nPossible types\ncallable Python object\npython object providing the\nwrite()\nmethod\nDefault value\nsys.stderr\nType of return value\nDepends on the selected feature strength calculation method:\nPredictionValuesChange, LossFunctionChange or PredictionValuesChange with the \nprettified\nparameter set to \nFalse\n: a list of length\n[n_features]\nwith float feature importances values for each feature\nPredictionValuesChange or LossFunctionChange with the \nprettified\nparameter set to \nTrue\n: a list of length\n[n_features]\nwith\n(feature_id (string), feature_importance (float))\npairs, sorted by feature importance values in descending order\nShapValues: np.array of shape \n(n_objects, n_features + 1)\nwith float ShapValues for each\n(object, feature)\nInteraction: list of length [ n_features] of three element lists of\n(first_feature_index, second_feature_index, interaction_score (float))","attrs_markdown":"[![Logo icon](https:\/\/yastatic.net\/s3\/locdoc\/daas-static\/catboost\/71b237a322eec6f2889af0dae2a9c549.svg)](https:\/\/catboost.ai\/ \"CatBoost\")\n\n- Installation\n  - [Overview](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/installation)\n  - Python package installation\n  - CatBoost for Apache Spark installation\n  - R package installation\n  - Command-line version binary\n  - Build from source\n- Key Features\n- Training parameters\n- Python package\n  - [Quick start](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-quickstart)\n  - CatBoost\n  - CatBoostClassifier\n    - [Overview](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier)\n    - [fit](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_fit)\n    - [predict](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_predict)\n    - [predict\\_proba](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_predict_proba)\n    - [Attributes](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_attributes)\n    - [calc\\_leaf\\_indexes](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboost_calc_leaf_indexes)\n    - [calc\\_feature\\_statistics](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_calc_feature_statistics)\n    - [compare](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_modelcompare)\n    - [copy](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_copy)\n    - [eval\\_metrics](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_eval-metrics)\n    - [get\\_all\\_params](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_all_params)\n    - [get\\_best\\_iteration](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_best_iteration)\n    - [get\\_best\\_score](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_best_score)\n    - [get\\_borders](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_borders)\n    - [get\\_evals\\_result](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_evals_result)\n    - [get\\_feature\\_importance](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance)\n    - [get\\_metadata](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_metadata)\n    - [get\\_object\\_importance](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_object_importance)\n    - [get\\_param](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_param)\n    - [get\\_params](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_params)\n    - [get\\_probability\\_threshold](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_probability_threshold)\n    - [get\\_scale\\_and\\_bias](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_scale_and_bias)\n    - [get\\_test\\_eval](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_test_eval)\n    - [grid\\_search](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_grid_search)\n    - [is\\_fitted](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_is_fitted)\n    - [load\\_model](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_load_model)\n    - [plot\\_predictions](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_plot_predictions)\n    - [plot\\_tree](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_plot_tree)\n    - [randomized\\_search](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_randomized_search)\n    - [save\\_borders](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_save_borders)\n    - [save\\_model](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_save_model)\n    - [score](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_score)\n    - [select\\_features](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_select_features)\n    - [set\\_feature\\_names](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_set_feature_names)\n    - [set\\_params](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_set_params)\n    - [set\\_probability\\_threshold](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_set_probability_threshold)\n    - [set\\_scale\\_and\\_bias](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_set_scale_and_bias)\n    - [shrink](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_shrink)\n    - [staged\\_predict](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_staged_predict)\n    - [staged\\_predict\\_proba](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_staged_predict_proba)\n  - CatBoostRanker\n  - CatBoostRegressor\n  - [cv](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_cv)\n  - datasets\n  - FeaturesData\n  - MetricVisualizer\n  - Pool\n  - [sample\\_gaussian\\_process](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_sample_gaussian_process)\n  - [sum\\_models](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_sum_models)\n  - [to\\_classifier](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_to_classifier)\n  - [to\\_regressor](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_to_regressor)\n  - [train](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_train)\n  - Text processing\n  - utils\n  - [Usage examples](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-usages-examples)\n- CatBoost for Apache Spark\n- R package\n- Command-line version\n- Applying models\n- Objectives and metrics\n- Model analysis\n- Data format description\n- [Parameter tuning](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/parameter-tuning)\n- [Speeding up the training](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/speed-up-training)\n- Data visualization\n- Algorithm details\n- [FAQ](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/faq)\n- Educational materials\n- [Development and contributions](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/development-and-contributions)\n- [Contacts](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/contacts)\n\nget\\_feature\\_importance\n\n## In this article:\n- [Method call format](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#call-format)\n- [Parameters](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#parameters)\n- [data](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#data)\n- [reference\\_data](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#reference_data)\n- [type](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#type)\n- [prettified](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#prettified)\n- [thread\\_count](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#thread_count)\n- [verbose](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#verbose)\n- [log\\_cout](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#log_cout)\n- [log\\_cerr](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#log_cerr)\n- [Type of return value](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#output-format)\n\n1. Python package\n2. [CatBoostClassifier](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier)\n3. get\\_feature\\_importance\n\n# get\\_feature\\_importance\n- [Method call format](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#call-format)\n- [Parameters](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#parameters)\n  - [data](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#data)\n  - [reference\\_data](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#reference_data)\n  - [type](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#type)\n  - [prettified](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#prettified)\n  - [thread\\_count](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#thread_count)\n  - [verbose](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#verbose)\n  - [log\\_cout](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#log_cout)\n  - [log\\_cerr](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#log_cerr)\n- [Type of return value](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_feature_importance#output-format)\n\nCalculate and return the [feature importances](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/fstr).\n\n## Method call format\n```\nget_feature_importance(data=None,\n                       reference_data=None,\n                       type=EFstrType.FeatureImportance,\n                       prettified=False,\n                       thread_count=-1,\n                       verbose=False,\n                       log_cout=sys.stdout,\n                       log_cerr=sys.stderr)\n```\n## Parameters\n### data\n#### Description\nThe dataset for feature importance calculation.\n\nThe required dataset depends on the selected feature importance calculation type (specified in the `type` parameter):\n\n- PredictionValuesChange — Either None or the same dataset that was used for training if the model does not contain information regarding the weight of leaves. All models trained with CatBoost version 0.9 or higher contain leaf weight information by default.\n- LossFunctionChange — Any dataset. Feature importances are calculated on a subset for large datasets.\n- PredictionDiff — A list of object pairs.\n\n**Possible types**\n\ncatboost.Pool\n\n**Default value**\n\nRequired parameter for the LossFunctionChange and ShapValues type of feature importances and in case the model does not contain information regarding the weight of leaves.\n\nNone otherwise.\n\n### reference\\_data\n#### Description\nReference data for Independent Tree SHAP values from [Explainable AI for Trees: From Local Explanations to Global Understanding](https:\/\/arxiv.org\/abs\/1905.04610v1). If `type` is [`ShapValues`](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/shap-values) and `reference_data` is not `None`, then Independent Tree SHAP values are calculated.\n\n**Possible types**\n\ncatboost.Pool\n\n**Default value**\n\nNone\n\n### type\n*Alias:*`fstr_type` (deprecated, use `type` instead)\n\n#### Description\nThe type of feature importance to calculate.\n\nPossible values:\n\n- FeatureImportance: Equal to [PredictionValuesChange](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/fstr#regular-feature-importance) for non-ranking metrics and [LossFunctionChange](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/fstr#regular-feature-importances__lossfunctionchange) for ranking metrics (the value is determined automatically).\n- [ShapValues](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/shap-values): A vector v v v with contributions of each feature to the prediction for every input object and the expected value of the model prediction for the object (average prediction given no knowledge about the object).\n- [Interaction](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/feature-interaction#feature-interaction-strength): The value of the feature interaction strength for each pair of features.\n- PredictionDiff: A vector with contributions of each feature to the RawFormulaVal difference for each pair of objects.\n\n**Possible types**\n\n- string\n- [EFStrType](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-efstr-type__desc)\n\nNote\n\nIt is recommended to use EFStrType for this parameter.\n\n**Default value**\n\nFeatureImportance\n\n### prettified\n#### Description\nReturn the feature importances as a list of the following pairs sorted by feature importance:\n```\n(feature_id, feature importance)\n```\nShould be used if one of the following values of the `type`parameter is selected:\n\n- PredictionValuesChange\n- PredictionValuesChange\n\n**Possible types**\n\nbool\n\n**Default value**\n\nFalse\n\n### thread\\_count\n#### Description\nThe number of threads to use for operation.\n\nOptimizes the speed of execution. This parameter doesn't affect results.\n\n**Possible types**\n\nint\n\n**Default value**\n\n\\-1 (the number of threads is equal to the number of processor cores)\n\n### verbose\n#### Description\nThe purpose of this parameter depends on the type of the given value:\n\n- bool — Output progress to stdout.\n  Works with the [ShapValues](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/shap-values) type of feature importance calculation.\n- int — The logging period.\n\n**Possible types**\n\n- bool\n- int\n\n**Default value**\n\nFalse\n\n### log\\_cout\nOutput stream or callback for logging.\n\n**Possible types**\n\n- callable Python object\n- python object providing the `write()` method\n\n**Default value**\n\nsys.stdout\n\n### log\\_cerr\nError stream or callback for logging.\n\n**Possible types**\n\n- callable Python object\n- python object providing the `write()` method\n\n**Default value**\n\nsys.stderr\n\n## Type of return value\nDepends on the selected feature strength calculation method:\n\n- PredictionValuesChange, LossFunctionChange or PredictionValuesChange with the `prettified` parameter set to \"False\": a list of length `[n_features]` with float feature importances values for each feature\n- PredictionValuesChange or LossFunctionChange with the `prettified` parameter set to \"True\": a list of length `[n_features]` with `(feature_id (string), feature_importance (float))` pairs, sorted by feature importance values in descending order\n- ShapValues: np.array of shape `(n_objects, n_features + 1)` with float ShapValues for each `(object, feature)`\n- Interaction: list of length \\[ n\\_features\\] of three element lists of `(first_feature_index, second_feature_index, interaction_score (float))`\n\n### Was the article helpful?\nYes\n\nNo\n\nPrevious\n\n[get\\_evals\\_result](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_get_evals_result)\n\nNext\n\n[get\\_metadata](https:\/\/catboost.ai\/docs\/en\/concepts\/en\/concepts\/python-reference_catboostclassifier_metadata)\n\n![](https:\/\/mc.yandex.ru\/watch\/60763294)","attrs_readable_markdown":"Calculate and return the [feature importances](https:\/\/catboost.ai\/docs\/en\/concepts\/fstr).\n\n## Method call format\n```\nget_feature_importance(data=None,\n                       reference_data=None,\n                       type=EFstrType.FeatureImportance,\n                       prettified=False,\n                       thread_count=-1,\n                       verbose=False,\n                       log_cout=sys.stdout,\n                       log_cerr=sys.stderr)\n```\n## Parameters\n### data\n#### Description\nThe dataset for feature importance calculation.\n\nThe required dataset depends on the selected feature importance calculation type (specified in the `type` parameter):\n\n- PredictionValuesChange — Either None or the same dataset that was used for training if the model does not contain information regarding the weight of leaves. All models trained with CatBoost version 0.9 or higher contain leaf weight information by default.\n- LossFunctionChange — Any dataset. Feature importances are calculated on a subset for large datasets.\n- PredictionDiff — A list of object pairs.\n\n**Possible types**\n\ncatboost.Pool\n\n**Default value**\n\nRequired parameter for the LossFunctionChange and ShapValues type of feature importances and in case the model does not contain information regarding the weight of leaves.\n\nNone otherwise.\n\n### reference\\_data\n#### Description\nReference data for Independent Tree SHAP values from [Explainable AI for Trees: From Local Explanations to Global Understanding](https:\/\/arxiv.org\/abs\/1905.04610v1). If `type` is [`ShapValues`](https:\/\/catboost.ai\/docs\/en\/concepts\/shap-values) and `reference_data` is not `None`, then Independent Tree SHAP values are calculated.\n\n**Possible types**\n\ncatboost.Pool\n\n**Default value**\n\nNone\n\n### type\n*Alias:*`fstr_type` (deprecated, use `type` instead)\n\n#### Description\nThe type of feature importance to calculate.\n\nPossible values:\n\n- FeatureImportance: Equal to [PredictionValuesChange](https:\/\/catboost.ai\/docs\/en\/concepts\/fstr#regular-feature-importance) for non-ranking metrics and [LossFunctionChange](https:\/\/catboost.ai\/docs\/en\/concepts\/fstr#regular-feature-importances__lossfunctionchange) for ranking metrics (the value is determined automatically).\n- [ShapValues](https:\/\/catboost.ai\/docs\/en\/concepts\/shap-values): A vector v v with contributions of each feature to the prediction for every input object and the expected value of the model prediction for the object (average prediction given no knowledge about the object).\n- [Interaction](https:\/\/catboost.ai\/docs\/en\/concepts\/feature-interaction#feature-interaction-strength): The value of the feature interaction strength for each pair of features.\n- PredictionDiff: A vector with contributions of each feature to the RawFormulaVal difference for each pair of objects.\n\n**Possible types**\n\n- string\n- [EFStrType](https:\/\/catboost.ai\/docs\/en\/concepts\/python-efstr-type__desc)\n\nNote\n\nIt is recommended to use EFStrType for this parameter.\n\n**Default value**\n\nFeatureImportance\n\n### prettified\n#### Description\nReturn the feature importances as a list of the following pairs sorted by feature importance:\n```\n(feature_id, feature importance)\n```\nShould be used if one of the following values of the `type`parameter is selected:\n\n- PredictionValuesChange\n- PredictionValuesChange\n\n**Possible types**\n\nbool\n\n**Default value**\n\nFalse\n\n### thread\\_count\n#### Description\nThe number of threads to use for operation.\n\nOptimizes the speed of execution. This parameter doesn't affect results.\n\n**Possible types**\n\nint\n\n**Default value**\n\n\\-1 (the number of threads is equal to the number of processor cores)\n\n### verbose\n#### Description\nThe purpose of this parameter depends on the type of the given value:\n\n- bool — Output progress to stdout.\n  Works with the [ShapValues](https:\/\/catboost.ai\/docs\/en\/concepts\/shap-values) type of feature importance calculation.\n- int — The logging period.\n\n**Possible types**\n\n- bool\n- int\n\n**Default value**\n\nFalse\n\n### log\\_cout\nOutput stream or callback for logging.\n\n**Possible types**\n\n- callable Python object\n- python object providing the `write()` method\n\n**Default value**\n\nsys.stdout\n\n### log\\_cerr\nError stream or callback for logging.\n\n**Possible types**\n\n- callable Python object\n- python object providing the `write()` method\n\n**Default value**\n\nsys.stderr\n\n## Type of return value\nDepends on the selected feature strength calculation method:\n\n- PredictionValuesChange, LossFunctionChange or PredictionValuesChange with the `prettified` parameter set to \"False\": a list of length `[n_features]` with float feature importances values for each feature\n- PredictionValuesChange or LossFunctionChange with the `prettified` parameter set to \"True\": a list of length `[n_features]` with `(feature_id (string), feature_importance (float))` pairs, sorted by feature importance values in descending order\n- ShapValues: np.array of shape `(n_objects, n_features + 1)` with float ShapValues for each `(object, feature)`\n- Interaction: list of length \\[ n\\_features\\] of three element lists of `(first_feature_index, second_feature_index, interaction_score (float))`","meta_canonical":null}

3. Robots.txt Check

Query:

Response:

4. Spam/Ban Check

Query:

Response:

5. Seen Status Check

ℹ️ Skipped - page is already crawled

📄

INDEXABLE

✅

CRAWLED

4 days ago

🤖

ROBOTS ALLOWED

Page Info Filters

Filter	Status	Condition	Details
HTTP status	PASS	`download_http_code = 200`	HTTP 200
Age cutoff	PASS	`download_stamp > now() - 6 MONTH`	0.2 months ago
History drop	PASS	`isNull(history_drop_reason)`	No drop reason
Spam/ban	PASS	`fh_dont_index != 1 AND ml_spam_score = 0`	ml_spam_score=0
Canonical	PASS	`meta_canonical IS NULL OR = '' OR = src_unparsed`	Not set

Page Details

Property	Value
URL	https://catboost.ai/docs/en/concepts/python-reference_catboostclassifier_get_feature_importance
Last Crawled	2026-04-04 06:51:47 (4 days ago)
First Indexed	2024-11-18 16:12:15 (1 year ago)
HTTP Status Code	200
Meta Title	get_feature_importance \| CatBoost
Meta Description	Calculate and return the feature importances. Method call format.
Meta Canonical	null
Boilerpipe Text	Calculate and return the feature importances . Method call format get_feature_importance(data= None , reference_data= None , type =EFstrType.FeatureImportance, prettified= False , thread_count=- 1 , verbose= False , log_cout=sys.stdout, log_cerr=sys.stderr) Parameters data Description The dataset for feature importance calculation. The required dataset depends on the selected feature importance calculation type (specified in the type parameter): PredictionValuesChange — Either None or the same dataset that was used for training if the model does not contain information regarding the weight of leaves. All models trained with CatBoost version 0.9 or higher contain leaf weight information by default. LossFunctionChange — Any dataset. Feature importances are calculated on a subset for large datasets. PredictionDiff — A list of object pairs. Possible types catboost.Pool Default value Required parameter for the LossFunctionChange and ShapValues type of feature importances and in case the model does not contain information regarding the weight of leaves. None otherwise. reference_data Description Reference data for Independent Tree SHAP values from Explainable AI for Trees: From Local Explanations to Global Understanding . If type is ShapValues and reference_data is not None , then Independent Tree SHAP values are calculated. Possible types catboost.Pool Default value None type Alias: fstr_type (deprecated, use type instead) Description The type of feature importance to calculate. Possible values: FeatureImportance: Equal to PredictionValuesChange for non-ranking metrics and LossFunctionChange for ranking metrics (the value is determined automatically). ShapValues : A vector v v with contributions of each feature to the prediction for every input object and the expected value of the model prediction for the object (average prediction given no knowledge about the object). Interaction : The value of the feature interaction strength for each pair of features. PredictionDiff: A vector with contributions of each feature to the RawFormulaVal difference for each pair of objects. Possible types string EFStrType Note It is recommended to use EFStrType for this parameter. Default value FeatureImportance prettified Description Return the feature importances as a list of the following pairs sorted by feature importance: (feature_id, feature importance) Should be used if one of the following values of the type parameter is selected: PredictionValuesChange PredictionValuesChange Possible types bool Default value False thread_count Description The number of threads to use for operation. Optimizes the speed of execution. This parameter doesn't affect results. Possible types int Default value -1 (the number of threads is equal to the number of processor cores) verbose Description The purpose of this parameter depends on the type of the given value: bool — Output progress to stdout. Works with the ShapValues type of feature importance calculation. int — The logging period. Possible types bool int Default value False log_cout Output stream or callback for logging. Possible types callable Python object python object providing the write() method Default value sys.stdout log_cerr Error stream or callback for logging. Possible types callable Python object python object providing the write() method Default value sys.stderr Type of return value Depends on the selected feature strength calculation method: PredictionValuesChange, LossFunctionChange or PredictionValuesChange with the prettified parameter set to False : a list of length [n_features] with float feature importances values for each feature PredictionValuesChange or LossFunctionChange with the prettified parameter set to True : a list of length [n_features] with (feature_id (string), feature_importance (float)) pairs, sorted by feature importance values in descending order ShapValues: np.array of shape (n_objects, n_features + 1) with float ShapValues for each (object, feature) Interaction: list of length [ n_features] of three element lists of (first_feature_index, second_feature_index, interaction_score (float))
Markdown	[![Logo icon](https://yastatic.net/s3/locdoc/daas-static/catboost/71b237a322eec6f2889af0dae2a9c549.svg)](https://catboost.ai/ "CatBoost") - Installation - [Overview](https://catboost.ai/docs/en/concepts/en/concepts/installation) - Python package installation - CatBoost for Apache Spark installation - R package installation - Command-line version binary - Build from source - Key Features - Training parameters - Python package - [Quick start](https://catboost.ai/docs/en/concepts/en/concepts/python-quickstart) - CatBoost - CatBoostClassifier - [Overview](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier) - [fit](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_fit) - [predict](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_predict) - [predict\_proba](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_predict_proba) - [Attributes](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_attributes) - [calc\_leaf\_indexes](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboost_calc_leaf_indexes) - [calc\_feature\_statistics](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_calc_feature_statistics) - [compare](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_modelcompare) - [copy](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_copy) - [eval\_metrics](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_eval-metrics) - [get\_all\_params](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_all_params) - [get\_best\_iteration](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_best_iteration) - [get\_best\_score](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_best_score) - [get\_borders](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_borders) - [get\_evals\_result](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_evals_result) - [get\_feature\_importance](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance) - [get\_metadata](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_metadata) - [get\_object\_importance](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_object_importance) - [get\_param](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_param) - [get\_params](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_params) - [get\_probability\_threshold](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_probability_threshold) - [get\_scale\_and\_bias](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_scale_and_bias) - [get\_test\_eval](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_test_eval) - [grid\_search](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_grid_search) - [is\_fitted](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_is_fitted) - [load\_model](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_load_model) - [plot\_predictions](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_plot_predictions) - [plot\_tree](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_plot_tree) - [randomized\_search](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_randomized_search) - [save\_borders](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_save_borders) - [save\_model](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_save_model) - [score](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_score) - [select\_features](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_select_features) - [set\_feature\_names](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_set_feature_names) - [set\_params](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_set_params) - [set\_probability\_threshold](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_set_probability_threshold) - [set\_scale\_and\_bias](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_set_scale_and_bias) - [shrink](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_shrink) - [staged\_predict](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_staged_predict) - [staged\_predict\_proba](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_staged_predict_proba) - CatBoostRanker - CatBoostRegressor - [cv](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_cv) - datasets - FeaturesData - MetricVisualizer - Pool - [sample\_gaussian\_process](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_sample_gaussian_process) - [sum\_models](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_sum_models) - [to\_classifier](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_to_classifier) - [to\_regressor](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_to_regressor) - [train](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_train) - Text processing - utils - [Usage examples](https://catboost.ai/docs/en/concepts/en/concepts/python-usages-examples) - CatBoost for Apache Spark - R package - Command-line version - Applying models - Objectives and metrics - Model analysis - Data format description - [Parameter tuning](https://catboost.ai/docs/en/concepts/en/concepts/parameter-tuning) - [Speeding up the training](https://catboost.ai/docs/en/concepts/en/concepts/speed-up-training) - Data visualization - Algorithm details - [FAQ](https://catboost.ai/docs/en/concepts/en/concepts/faq) - Educational materials - [Development and contributions](https://catboost.ai/docs/en/concepts/en/concepts/development-and-contributions) - [Contacts](https://catboost.ai/docs/en/concepts/en/concepts/contacts) get\_feature\_importance ## In this article: - [Method call format](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#call-format) - [Parameters](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#parameters) - [data](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#data) - [reference\_data](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#reference_data) - [type](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#type) - [prettified](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#prettified) - [thread\_count](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#thread_count) - [verbose](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#verbose) - [log\_cout](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#log_cout) - [log\_cerr](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#log_cerr) - [Type of return value](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#output-format) 1. Python package 2. [CatBoostClassifier](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier) 3. get\_feature\_importance # get\_feature\_importance - [Method call format](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#call-format) - [Parameters](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#parameters) - [data](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#data) - [reference\_data](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#reference_data) - [type](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#type) - [prettified](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#prettified) - [thread\_count](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#thread_count) - [verbose](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#verbose) - [log\_cout](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#log_cout) - [log\_cerr](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#log_cerr) - [Type of return value](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_feature_importance#output-format) Calculate and return the [feature importances](https://catboost.ai/docs/en/concepts/en/concepts/fstr). ## Method call format ``` get_feature_importance(data=None, reference_data=None, type=EFstrType.FeatureImportance, prettified=False, thread_count=-1, verbose=False, log_cout=sys.stdout, log_cerr=sys.stderr) ``` ## Parameters ### data #### Description The dataset for feature importance calculation. The required dataset depends on the selected feature importance calculation type (specified in the `type` parameter): - PredictionValuesChange — Either None or the same dataset that was used for training if the model does not contain information regarding the weight of leaves. All models trained with CatBoost version 0.9 or higher contain leaf weight information by default. - LossFunctionChange — Any dataset. Feature importances are calculated on a subset for large datasets. - PredictionDiff — A list of object pairs. Possible types catboost.Pool Default value Required parameter for the LossFunctionChange and ShapValues type of feature importances and in case the model does not contain information regarding the weight of leaves. None otherwise. ### reference\_data #### Description Reference data for Independent Tree SHAP values from [Explainable AI for Trees: From Local Explanations to Global Understanding](https://arxiv.org/abs/1905.04610v1). If `type` is [`ShapValues`](https://catboost.ai/docs/en/concepts/en/concepts/shap-values) and `reference_data` is not `None`, then Independent Tree SHAP values are calculated. Possible types catboost.Pool Default value None ### type Alias:`fstr_type` (deprecated, use `type` instead) #### Description The type of feature importance to calculate. Possible values: - FeatureImportance: Equal to [PredictionValuesChange](https://catboost.ai/docs/en/concepts/en/concepts/fstr#regular-feature-importance) for non-ranking metrics and [LossFunctionChange](https://catboost.ai/docs/en/concepts/en/concepts/fstr#regular-feature-importances__lossfunctionchange) for ranking metrics (the value is determined automatically). - [ShapValues](https://catboost.ai/docs/en/concepts/en/concepts/shap-values): A vector v v v with contributions of each feature to the prediction for every input object and the expected value of the model prediction for the object (average prediction given no knowledge about the object). - [Interaction](https://catboost.ai/docs/en/concepts/en/concepts/feature-interaction#feature-interaction-strength): The value of the feature interaction strength for each pair of features. - PredictionDiff: A vector with contributions of each feature to the RawFormulaVal difference for each pair of objects. Possible types - string - [EFStrType](https://catboost.ai/docs/en/concepts/en/concepts/python-efstr-type__desc) Note It is recommended to use EFStrType for this parameter. Default value FeatureImportance ### prettified #### Description Return the feature importances as a list of the following pairs sorted by feature importance: ``` (feature_id, feature importance) ``` Should be used if one of the following values of the `type`parameter is selected: - PredictionValuesChange - PredictionValuesChange Possible types bool Default value False ### thread\_count #### Description The number of threads to use for operation. Optimizes the speed of execution. This parameter doesn't affect results. Possible types int Default value \-1 (the number of threads is equal to the number of processor cores) ### verbose #### Description The purpose of this parameter depends on the type of the given value: - bool — Output progress to stdout. Works with the [ShapValues](https://catboost.ai/docs/en/concepts/en/concepts/shap-values) type of feature importance calculation. - int — The logging period. Possible types - bool - int Default value False ### log\_cout Output stream or callback for logging. Possible types - callable Python object - python object providing the `write()` method Default value sys.stdout ### log\_cerr Error stream or callback for logging. Possible types - callable Python object - python object providing the `write()` method Default value sys.stderr ## Type of return value Depends on the selected feature strength calculation method: - PredictionValuesChange, LossFunctionChange or PredictionValuesChange with the `prettified` parameter set to "False": a list of length `[n_features]` with float feature importances values for each feature - PredictionValuesChange or LossFunctionChange with the `prettified` parameter set to "True": a list of length `[n_features]` with `(feature_id (string), feature_importance (float))` pairs, sorted by feature importance values in descending order - ShapValues: np.array of shape `(n_objects, n_features + 1)` with float ShapValues for each `(object, feature)` - Interaction: list of length \[ n\_features\] of three element lists of `(first_feature_index, second_feature_index, interaction_score (float))` ### Was the article helpful? Yes No Previous [get\_evals\_result](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_get_evals_result) Next [get\_metadata](https://catboost.ai/docs/en/concepts/en/concepts/python-reference_catboostclassifier_metadata) ![](https://mc.yandex.ru/watch/60763294)
Readable Markdown	Calculate and return the [feature importances](https://catboost.ai/docs/en/concepts/fstr). ## Method call format ``` get_feature_importance(data=None, reference_data=None, type=EFstrType.FeatureImportance, prettified=False, thread_count=-1, verbose=False, log_cout=sys.stdout, log_cerr=sys.stderr) ``` ## Parameters ### data #### Description The dataset for feature importance calculation. The required dataset depends on the selected feature importance calculation type (specified in the `type` parameter): - PredictionValuesChange — Either None or the same dataset that was used for training if the model does not contain information regarding the weight of leaves. All models trained with CatBoost version 0.9 or higher contain leaf weight information by default. - LossFunctionChange — Any dataset. Feature importances are calculated on a subset for large datasets. - PredictionDiff — A list of object pairs. Possible types catboost.Pool Default value Required parameter for the LossFunctionChange and ShapValues type of feature importances and in case the model does not contain information regarding the weight of leaves. None otherwise. ### reference\_data #### Description Reference data for Independent Tree SHAP values from [Explainable AI for Trees: From Local Explanations to Global Understanding](https://arxiv.org/abs/1905.04610v1). If `type` is [`ShapValues`](https://catboost.ai/docs/en/concepts/shap-values) and `reference_data` is not `None`, then Independent Tree SHAP values are calculated. Possible types catboost.Pool Default value None ### type Alias:`fstr_type` (deprecated, use `type` instead) #### Description The type of feature importance to calculate. Possible values: - FeatureImportance: Equal to [PredictionValuesChange](https://catboost.ai/docs/en/concepts/fstr#regular-feature-importance) for non-ranking metrics and [LossFunctionChange](https://catboost.ai/docs/en/concepts/fstr#regular-feature-importances__lossfunctionchange) for ranking metrics (the value is determined automatically). - [ShapValues](https://catboost.ai/docs/en/concepts/shap-values): A vector v v with contributions of each feature to the prediction for every input object and the expected value of the model prediction for the object (average prediction given no knowledge about the object). - [Interaction](https://catboost.ai/docs/en/concepts/feature-interaction#feature-interaction-strength): The value of the feature interaction strength for each pair of features. - PredictionDiff: A vector with contributions of each feature to the RawFormulaVal difference for each pair of objects. Possible types - string - [EFStrType](https://catboost.ai/docs/en/concepts/python-efstr-type__desc) Note It is recommended to use EFStrType for this parameter. Default value FeatureImportance ### prettified #### Description Return the feature importances as a list of the following pairs sorted by feature importance: ``` (feature_id, feature importance) ``` Should be used if one of the following values of the `type`parameter is selected: - PredictionValuesChange - PredictionValuesChange Possible types bool Default value False ### thread\_count #### Description The number of threads to use for operation. Optimizes the speed of execution. This parameter doesn't affect results. Possible types int Default value \-1 (the number of threads is equal to the number of processor cores) ### verbose #### Description The purpose of this parameter depends on the type of the given value: - bool — Output progress to stdout. Works with the [ShapValues](https://catboost.ai/docs/en/concepts/shap-values) type of feature importance calculation. - int — The logging period. Possible types - bool - int Default value False ### log\_cout Output stream or callback for logging. Possible types - callable Python object - python object providing the `write()` method Default value sys.stdout ### log\_cerr Error stream or callback for logging. Possible types - callable Python object - python object providing the `write()` method Default value sys.stderr ## Type of return value Depends on the selected feature strength calculation method: - PredictionValuesChange, LossFunctionChange or PredictionValuesChange with the `prettified` parameter set to "False": a list of length `[n_features]` with float feature importances values for each feature - PredictionValuesChange or LossFunctionChange with the `prettified` parameter set to "True": a list of length `[n_features]` with `(feature_id (string), feature_importance (float))` pairs, sorted by feature importance values in descending order - ShapValues: np.array of shape `(n_objects, n_features + 1)` with float ShapValues for each `(object, feature)` - Interaction: list of length \[ n\_features\] of three element lists of `(first_feature_index, second_feature_index, interaction_score (float))`
Shard	169 (laksa)
Root Hash	17435841955170310369
Unparsed URL	ai,catboost!/docs/en/concepts/python-reference_catboostclassifier_get_feature_importance s443