Merge remote-tracking branch 'origin/ts_forecasting_advanced' into ts…

…_forecasting_advanced
aimclub · May 17, 2024 · 30bbe4b · 30bbe4b
2 parents 5001929 + 3bb3246
commit 30bbe4b
Show file tree

Hide file tree

Showing 4 changed files with 25 additions and 21 deletions.
diff --git a/...xample/api_example/time_series/ts_forecasting/m4_analiysis/ts_forecasting_m4_benchmark.py b/...xample/api_example/time_series/ts_forecasting/m4_analiysis/ts_forecasting_m4_benchmark.py
@@ -42,8 +42,8 @@
             api_config.update(output_folder=os.path.join(
                 PROJECT_PATH, 'results_of_experiments', dataset_name))
             n_beats_forecast, n_beats_metrics, \
-            autogluon_forecast, autogluon_metrics = compare_forecast_with_sota(dataset_name=dataset_name,
-                                                                               horizon=horizon)
+                autogluon_forecast, autogluon_metrics = compare_forecast_with_sota(dataset_name=dataset_name,
+                                                                                   horizon=horizon)
             model, labels, metrics, target = industrial_forecasting_modelling_loop(dataset_name=dataset_name,
                                                                                    benchmark=benchmark,
                                                                                    horizon=horizon,

diff --git a/examples/automl_example/api_example/time_series/ts_forecasting/ts_forecasting_example.py b/examples/automl_example/api_example/time_series/ts_forecasting/ts_forecasting_example.py
@@ -6,12 +6,13 @@
     dataset_name = {'benchmark': 'M4',
                     'dataset': 'D3257'}
     finetune = False
-    initial_assumptions = {'nbeats': PipelineBuilder().add_node('nbeats_model'),
-                           'industiral': PipelineBuilder().add_node(
-                               'eigen_basis',
-                               params={
-                                   'low_rank_approximation': False,
-                                   'rank_regularization': 'explained_dispersion'}).add_node('ar')}
+    initial_assumptions = {
+        'nbeats': PipelineBuilder().add_node('nbeats_model'),
+        'industiral': PipelineBuilder().add_node(
+            'eigen_basis',
+            params={
+                'low_rank_approximation': False,
+                'rank_regularization': 'explained_dispersion'}).add_node('ar')}
     for assumption in initial_assumptions.keys():
         api_config = dict(problem='ts_forecasting',
                           metric='rmse',

diff --git a/fedot_ind/core/metrics/metrics_implementation.py b/fedot_ind/core/metrics/metrics_implementation.py
@@ -266,12 +266,12 @@ def calculate_classification_metric(
     probs,
     rounding_order=3,
     metric_names=(
-            'f1',
-            # 'roc_auc',
-            'accuracy')):
+        'f1',
+        # 'roc_auc',
+        'accuracy')):
     metric_dict = {'accuracy': Accuracy,
                    'f1': F1,
-                   #'roc_auc': ROCAUC,
+                   # 'roc_auc': ROCAUC,
                    'precision': Precision,
                    'logloss': Logloss}
 

diff --git a/fedot_ind/tools/example_utils.py b/fedot_ind/tools/example_utils.py
@@ -93,18 +93,21 @@ def industrial_common_modelling_loop(
         finetune: bool = False,
         api_config: dict = None,
         metric_names: tuple = (
-                'r2',
-                'rmse',
-                'mae')):
+            'r2',
+            'rmse',
+            'mae')):
     industrial = FedotIndustrial(**api_config)
     if api_config['problem'] == 'ts_forecasting':
-        train_data, _ = DataLoader(dataset_name=dataset_name['dataset']).load_forecast_data(
+        train_data, _ = DataLoader(
+            dataset_name=dataset_name['dataset']).load_forecast_data(
             dataset_name['benchmark'])
-        target = train_data.values[-api_config['task_params']['forecast_length']:].flatten()
+        target = train_data.values[-api_config['task_params']
+                                   ['forecast_length']:].flatten()
         train_data = (train_data, target)
         test_data = train_data
     else:
-        train_data, test_data = DataLoader(dataset_name=dataset_name).load_data()
+        train_data, test_data = DataLoader(
+            dataset_name=dataset_name).load_data()
     if finetune:
         industrial.finetune(train_data, tuning_params={
             'tuning_timeout': api_config['timeout']})
@@ -139,11 +142,11 @@ def create_comprasion_df(df, metric: str = 'rmse'):
     df_full = df_full[df_full['Unnamed: 0'] == metric]
     df_full = df_full.drop('Unnamed: 0', axis=1)
     df_full['Difference_industrial_All'] = (
-            df_full.iloc[:, 1:3].min(axis=1) - df_full['industrial'])
+        df_full.iloc[:, 1:3].min(axis=1) - df_full['industrial'])
     df_full['Difference_industrial_AG'] = (
-            df_full.iloc[:, 1:2].min(axis=1) - df_full['industrial'])
+        df_full.iloc[:, 1:2].min(axis=1) - df_full['industrial'])
     df_full['Difference_industrial_NBEATS'] = (
-            df_full.iloc[:, 2:3].min(axis=1) - df_full['industrial'])
+        df_full.iloc[:, 2:3].min(axis=1) - df_full['industrial'])
     df_full['industrial_Wins_All'] = df_full.apply(
         lambda row: 'Win' if row.loc['Difference_industrial_All'] > 0 else 'Loose', axis=1)
     df_full['industrial_Wins_AG'] = df_full.apply(