-
Notifications
You must be signed in to change notification settings - Fork 1.6k
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Components - XGBoost - Added the Cross_validation_for_regression comp…
…onent (#4244)
- Loading branch information
Showing
2 changed files
with
339 additions
and
0 deletions.
There are no files selected for viewing
67 changes: 67 additions & 0 deletions
67
components/XGBoost/Cross_validation_for_regression/from_CSV/component.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,67 @@ | ||
from collections import OrderedDict | ||
from kfp import components | ||
|
||
|
||
split_table_into_folds_op = components.load_component_from_url('https://raw.githubusercontent.com/kubeflow/pipelines/e9b4b29b22a5120daf95b581b0392cd461a906f0/components/dataset_manipulation/split_data_into_folds/in_CSV/component.yaml') | ||
xgboost_train_on_csv_op = components.load_component_from_url('https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Train/component.yaml') | ||
xgboost_predict_on_csv_op = components.load_component_from_url('https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Predict/component.yaml') | ||
pandas_transform_csv_op = components.load_component_from_url('https://raw.githubusercontent.com/kubeflow/pipelines/6162d55998b176b50267d351241100bb0ee715bc/components/pandas/Transform_DataFrame/in_CSV_format/component.yaml') | ||
drop_header_op = components.load_component_from_url('https://raw.githubusercontent.com/kubeflow/pipelines/02c9638287468c849632cf9f7885b51de4c66f86/components/tables/Remove_header/component.yaml') | ||
calculate_regression_metrics_from_csv_op = components.load_component_from_url('https://raw.githubusercontent.com/kubeflow/pipelines/7da1ac9464b4b3e7d95919faa2f1107a9635b7e4/components/ml_metrics/Calculate_regression_metrics/from_CSV/component.yaml') | ||
aggregate_regression_metrics_op = components.load_component_from_url('https://raw.githubusercontent.com/kubeflow/pipelines/7ea9363fe201918d419fecdc00d1275e657ff712/components/ml_metrics/Aggregate_regression_metrics/component.yaml') | ||
|
||
|
||
def xgboost_5_fold_cross_validation_for_regression( | ||
data: 'CSV', | ||
label_column: int = 0, | ||
objective: str = 'reg:squarederror', | ||
num_iterations: int = 200, | ||
): | ||
folds = split_table_into_folds_op(data).outputs | ||
|
||
fold_metrics = {} | ||
for i in range(1, 6): | ||
training_data = folds['train_' + str(i)] | ||
testing_data = folds['test_' + str(i)] | ||
model = xgboost_train_on_csv_op( | ||
training_data=training_data, | ||
label_column=label_column, | ||
objective=objective, | ||
num_iterations=num_iterations, | ||
).outputs['model'] | ||
|
||
predictions = xgboost_predict_on_csv_op( | ||
data=testing_data, | ||
model=model, | ||
label_column=label_column, | ||
).output | ||
|
||
true_values_table = pandas_transform_csv_op( | ||
table=testing_data, | ||
transform_code='df = df[["tips"]]', | ||
).output | ||
|
||
true_values = drop_header_op(true_values_table).output | ||
|
||
metrics = calculate_regression_metrics_from_csv_op( | ||
true_values=true_values, | ||
predicted_values=predictions, | ||
).outputs['metrics'] | ||
|
||
fold_metrics['metrics_' + str(i)] = metrics | ||
|
||
aggregated_metrics_task = aggregate_regression_metrics_op(**fold_metrics) | ||
|
||
return OrderedDict([ | ||
('mean_absolute_error', aggregated_metrics_task.outputs['mean_absolute_error']), | ||
('mean_squared_error', aggregated_metrics_task.outputs['mean_squared_error']), | ||
('root_mean_squared_error', aggregated_metrics_task.outputs['root_mean_squared_error']), | ||
('metrics', aggregated_metrics_task.outputs['metrics']), | ||
]) | ||
|
||
|
||
if __name__ == '__main__': | ||
xgboost_5_fold_cross_validation_for_regression_op = components.create_graph_component_from_pipeline_func( | ||
xgboost_5_fold_cross_validation_for_regression, | ||
output_component_file='component.yaml', | ||
) |
272 changes: 272 additions & 0 deletions
272
components/XGBoost/Cross_validation_for_regression/from_CSV/component.yaml
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,272 @@ | ||
name: Xgboost 5 fold cross validation for regression | ||
inputs: | ||
- {name: data, type: CSV} | ||
- {name: label_column, type: Integer, default: '0', optional: true} | ||
- {name: objective, type: String, default: 'reg:squarederror', optional: true} | ||
- {name: num_iterations, type: Integer, default: '200', optional: true} | ||
outputs: | ||
- {name: mean_absolute_error, type: Float} | ||
- {name: mean_squared_error, type: Float} | ||
- {name: root_mean_squared_error, type: Float} | ||
- {name: metrics, type: JsonObject} | ||
implementation: | ||
graph: | ||
tasks: | ||
Split table into folds: | ||
componentRef: {digest: 9956223bcecc7294ca1afac39b60ada4a935a571d817c3dfbf2ea4a211afe3d1, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/e9b4b29b22a5120daf95b581b0392cd461a906f0/components/dataset_manipulation/split_data_into_folds/in_CSV/component.yaml'} | ||
arguments: | ||
table: | ||
graphInput: {inputName: data} | ||
Xgboost train: | ||
componentRef: {digest: 09b80053da29f8f51575b42e5d2e8ad4b7bdcc92a02c3744e189b1f597006b38, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Train/component.yaml'} | ||
arguments: | ||
training_data: | ||
taskOutput: {outputName: train_1, taskId: Split table into folds, type: CSV} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
num_iterations: | ||
graphInput: {inputName: num_iterations} | ||
objective: | ||
graphInput: {inputName: objective} | ||
Xgboost predict: | ||
componentRef: {digest: ecdfaf32cff15b6abc3d0dd80365ce00577f1a19a058fbe201f515431cea1357, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Predict/component.yaml'} | ||
arguments: | ||
data: | ||
taskOutput: {outputName: test_1, taskId: Split table into folds, type: CSV} | ||
model: | ||
taskOutput: {outputName: model, taskId: Xgboost train, type: XGBoostModel} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
Pandas Transform DataFrame in CSV format: | ||
componentRef: {digest: 58dc88349157bf128021708c316ce4eb60bc1de0a5a7dd3af45fabac3276d510, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/6162d55998b176b50267d351241100bb0ee715bc/components/pandas/Transform_DataFrame/in_CSV_format/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: test_1, taskId: Split table into folds, type: CSV} | ||
transform_code: df = df[["tips"]] | ||
Remove header: | ||
componentRef: {digest: ba35ffea863855b956c3c50aefa0420ba3823949a6c059e6e3971cde960dc5a3, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/02c9638287468c849632cf9f7885b51de4c66f86/components/tables/Remove_header/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: transformed_table, taskId: Pandas Transform DataFrame | ||
in CSV format, type: CSV} | ||
Calculate regression metrics from csv: | ||
componentRef: {digest: e3ecbfeb18032820edfee4255e2fb6d15d15ed224e166519d5e528e12053a995, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/7da1ac9464b4b3e7d95919faa2f1107a9635b7e4/components/ml_metrics/Calculate_regression_metrics/from_CSV/component.yaml'} | ||
arguments: | ||
true_values: | ||
taskOutput: {outputName: table, taskId: Remove header} | ||
predicted_values: | ||
taskOutput: {outputName: predictions, taskId: Xgboost predict, type: Text} | ||
Xgboost train 2: | ||
componentRef: {digest: 09b80053da29f8f51575b42e5d2e8ad4b7bdcc92a02c3744e189b1f597006b38, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Train/component.yaml'} | ||
arguments: | ||
training_data: | ||
taskOutput: {outputName: train_2, taskId: Split table into folds, type: CSV} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
num_iterations: | ||
graphInput: {inputName: num_iterations} | ||
objective: | ||
graphInput: {inputName: objective} | ||
Xgboost predict 2: | ||
componentRef: {digest: ecdfaf32cff15b6abc3d0dd80365ce00577f1a19a058fbe201f515431cea1357, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Predict/component.yaml'} | ||
arguments: | ||
data: | ||
taskOutput: {outputName: test_2, taskId: Split table into folds, type: CSV} | ||
model: | ||
taskOutput: {outputName: model, taskId: Xgboost train 2, type: XGBoostModel} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
Pandas Transform DataFrame in CSV format 2: | ||
componentRef: {digest: 58dc88349157bf128021708c316ce4eb60bc1de0a5a7dd3af45fabac3276d510, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/6162d55998b176b50267d351241100bb0ee715bc/components/pandas/Transform_DataFrame/in_CSV_format/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: test_2, taskId: Split table into folds, type: CSV} | ||
transform_code: df = df[["tips"]] | ||
Remove header 2: | ||
componentRef: {digest: ba35ffea863855b956c3c50aefa0420ba3823949a6c059e6e3971cde960dc5a3, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/02c9638287468c849632cf9f7885b51de4c66f86/components/tables/Remove_header/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: transformed_table, taskId: Pandas Transform DataFrame | ||
in CSV format 2, type: CSV} | ||
Calculate regression metrics from csv 2: | ||
componentRef: {digest: e3ecbfeb18032820edfee4255e2fb6d15d15ed224e166519d5e528e12053a995, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/7da1ac9464b4b3e7d95919faa2f1107a9635b7e4/components/ml_metrics/Calculate_regression_metrics/from_CSV/component.yaml'} | ||
arguments: | ||
true_values: | ||
taskOutput: {outputName: table, taskId: Remove header 2} | ||
predicted_values: | ||
taskOutput: {outputName: predictions, taskId: Xgboost predict 2, type: Text} | ||
Xgboost train 3: | ||
componentRef: {digest: 09b80053da29f8f51575b42e5d2e8ad4b7bdcc92a02c3744e189b1f597006b38, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Train/component.yaml'} | ||
arguments: | ||
training_data: | ||
taskOutput: {outputName: train_3, taskId: Split table into folds, type: CSV} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
num_iterations: | ||
graphInput: {inputName: num_iterations} | ||
objective: | ||
graphInput: {inputName: objective} | ||
Xgboost predict 3: | ||
componentRef: {digest: ecdfaf32cff15b6abc3d0dd80365ce00577f1a19a058fbe201f515431cea1357, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Predict/component.yaml'} | ||
arguments: | ||
data: | ||
taskOutput: {outputName: test_3, taskId: Split table into folds, type: CSV} | ||
model: | ||
taskOutput: {outputName: model, taskId: Xgboost train 3, type: XGBoostModel} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
Pandas Transform DataFrame in CSV format 3: | ||
componentRef: {digest: 58dc88349157bf128021708c316ce4eb60bc1de0a5a7dd3af45fabac3276d510, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/6162d55998b176b50267d351241100bb0ee715bc/components/pandas/Transform_DataFrame/in_CSV_format/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: test_3, taskId: Split table into folds, type: CSV} | ||
transform_code: df = df[["tips"]] | ||
Remove header 3: | ||
componentRef: {digest: ba35ffea863855b956c3c50aefa0420ba3823949a6c059e6e3971cde960dc5a3, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/02c9638287468c849632cf9f7885b51de4c66f86/components/tables/Remove_header/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: transformed_table, taskId: Pandas Transform DataFrame | ||
in CSV format 3, type: CSV} | ||
Calculate regression metrics from csv 3: | ||
componentRef: {digest: e3ecbfeb18032820edfee4255e2fb6d15d15ed224e166519d5e528e12053a995, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/7da1ac9464b4b3e7d95919faa2f1107a9635b7e4/components/ml_metrics/Calculate_regression_metrics/from_CSV/component.yaml'} | ||
arguments: | ||
true_values: | ||
taskOutput: {outputName: table, taskId: Remove header 3} | ||
predicted_values: | ||
taskOutput: {outputName: predictions, taskId: Xgboost predict 3, type: Text} | ||
Xgboost train 4: | ||
componentRef: {digest: 09b80053da29f8f51575b42e5d2e8ad4b7bdcc92a02c3744e189b1f597006b38, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Train/component.yaml'} | ||
arguments: | ||
training_data: | ||
taskOutput: {outputName: train_4, taskId: Split table into folds, type: CSV} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
num_iterations: | ||
graphInput: {inputName: num_iterations} | ||
objective: | ||
graphInput: {inputName: objective} | ||
Xgboost predict 4: | ||
componentRef: {digest: ecdfaf32cff15b6abc3d0dd80365ce00577f1a19a058fbe201f515431cea1357, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Predict/component.yaml'} | ||
arguments: | ||
data: | ||
taskOutput: {outputName: test_4, taskId: Split table into folds, type: CSV} | ||
model: | ||
taskOutput: {outputName: model, taskId: Xgboost train 4, type: XGBoostModel} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
Pandas Transform DataFrame in CSV format 4: | ||
componentRef: {digest: 58dc88349157bf128021708c316ce4eb60bc1de0a5a7dd3af45fabac3276d510, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/6162d55998b176b50267d351241100bb0ee715bc/components/pandas/Transform_DataFrame/in_CSV_format/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: test_4, taskId: Split table into folds, type: CSV} | ||
transform_code: df = df[["tips"]] | ||
Remove header 4: | ||
componentRef: {digest: ba35ffea863855b956c3c50aefa0420ba3823949a6c059e6e3971cde960dc5a3, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/02c9638287468c849632cf9f7885b51de4c66f86/components/tables/Remove_header/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: transformed_table, taskId: Pandas Transform DataFrame | ||
in CSV format 4, type: CSV} | ||
Calculate regression metrics from csv 4: | ||
componentRef: {digest: e3ecbfeb18032820edfee4255e2fb6d15d15ed224e166519d5e528e12053a995, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/7da1ac9464b4b3e7d95919faa2f1107a9635b7e4/components/ml_metrics/Calculate_regression_metrics/from_CSV/component.yaml'} | ||
arguments: | ||
true_values: | ||
taskOutput: {outputName: table, taskId: Remove header 4} | ||
predicted_values: | ||
taskOutput: {outputName: predictions, taskId: Xgboost predict 4, type: Text} | ||
Xgboost train 5: | ||
componentRef: {digest: 09b80053da29f8f51575b42e5d2e8ad4b7bdcc92a02c3744e189b1f597006b38, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Train/component.yaml'} | ||
arguments: | ||
training_data: | ||
taskOutput: {outputName: train_5, taskId: Split table into folds, type: CSV} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
num_iterations: | ||
graphInput: {inputName: num_iterations} | ||
objective: | ||
graphInput: {inputName: objective} | ||
Xgboost predict 5: | ||
componentRef: {digest: ecdfaf32cff15b6abc3d0dd80365ce00577f1a19a058fbe201f515431cea1357, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/567c04c51ff00a1ee525b3458425b17adbe3df61/components/XGBoost/Predict/component.yaml'} | ||
arguments: | ||
data: | ||
taskOutput: {outputName: test_5, taskId: Split table into folds, type: CSV} | ||
model: | ||
taskOutput: {outputName: model, taskId: Xgboost train 5, type: XGBoostModel} | ||
label_column: | ||
graphInput: {inputName: label_column} | ||
Pandas Transform DataFrame in CSV format 5: | ||
componentRef: {digest: 58dc88349157bf128021708c316ce4eb60bc1de0a5a7dd3af45fabac3276d510, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/6162d55998b176b50267d351241100bb0ee715bc/components/pandas/Transform_DataFrame/in_CSV_format/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: test_5, taskId: Split table into folds, type: CSV} | ||
transform_code: df = df[["tips"]] | ||
Remove header 5: | ||
componentRef: {digest: ba35ffea863855b956c3c50aefa0420ba3823949a6c059e6e3971cde960dc5a3, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/02c9638287468c849632cf9f7885b51de4c66f86/components/tables/Remove_header/component.yaml'} | ||
arguments: | ||
table: | ||
taskOutput: {outputName: transformed_table, taskId: Pandas Transform DataFrame | ||
in CSV format 5, type: CSV} | ||
Calculate regression metrics from csv 5: | ||
componentRef: {digest: e3ecbfeb18032820edfee4255e2fb6d15d15ed224e166519d5e528e12053a995, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/7da1ac9464b4b3e7d95919faa2f1107a9635b7e4/components/ml_metrics/Calculate_regression_metrics/from_CSV/component.yaml'} | ||
arguments: | ||
true_values: | ||
taskOutput: {outputName: table, taskId: Remove header 5} | ||
predicted_values: | ||
taskOutput: {outputName: predictions, taskId: Xgboost predict 5, type: Text} | ||
Aggregate regression metrics from csv: | ||
componentRef: {digest: 3e128130521eff8d43764f3dcb037316cdd6490ad2878df5adef416f7c2f3c19, | ||
url: 'https://raw.githubusercontent.com/kubeflow/pipelines/7ea9363fe201918d419fecdc00d1275e657ff712/components/ml_metrics/Aggregate_regression_metrics/component.yaml'} | ||
arguments: | ||
metrics_1: | ||
taskOutput: {outputName: metrics, taskId: Calculate regression metrics | ||
from csv, type: JsonObject} | ||
metrics_2: | ||
taskOutput: {outputName: metrics, taskId: Calculate regression metrics | ||
from csv 2, type: JsonObject} | ||
metrics_3: | ||
taskOutput: {outputName: metrics, taskId: Calculate regression metrics | ||
from csv 3, type: JsonObject} | ||
metrics_4: | ||
taskOutput: {outputName: metrics, taskId: Calculate regression metrics | ||
from csv 4, type: JsonObject} | ||
metrics_5: | ||
taskOutput: {outputName: metrics, taskId: Calculate regression metrics | ||
from csv 5, type: JsonObject} | ||
outputValues: | ||
mean_absolute_error: | ||
taskOutput: {outputName: mean_absolute_error, taskId: Aggregate regression | ||
metrics from csv, type: Float} | ||
mean_squared_error: | ||
taskOutput: {outputName: mean_squared_error, taskId: Aggregate regression | ||
metrics from csv, type: Float} | ||
root_mean_squared_error: | ||
taskOutput: {outputName: root_mean_squared_error, taskId: Aggregate regression | ||
metrics from csv, type: Float} | ||
metrics: | ||
taskOutput: {outputName: metrics, taskId: Aggregate regression metrics from | ||
csv, type: JsonObject} |