[ENH] AutoETS, ARIMA, NaiveForecaster, Dataset loading updates, Differencing/Windowing/Train test transforms.

aeon/datasets/Final Dataset Selection.csv

@@ -0,0 +1,101 @@
+Dataset,Series,Category
+weather_dataset,T1,Weather
+weather_dataset,T2,Weather
+weather_dataset,T3,Weather
+weather_dataset,T4,Weather
+weather_dataset,T5,Weather
+solar_10_minutes_dataset,T1,Energy Production
+solar_10_minutes_dataset,T2,Energy Production
+solar_10_minutes_dataset,T3,Energy Production
+solar_10_minutes_dataset,T4,Energy Production
+solar_10_minutes_dataset,T5,Energy Production
+sunspot_dataset_without_missing_values,T1,Other
+wind_farms_minutely_dataset_without_missing_values,T1,Energy Production
+wind_farms_minutely_dataset_without_missing_values,T2,Energy Production
+wind_farms_minutely_dataset_without_missing_values,T3,Energy Production
+wind_farms_minutely_dataset_without_missing_values,T4,Energy Production
+wind_farms_minutely_dataset_without_missing_values,T5,Energy Production
+elecdemand_dataset,T1,Energy Demand
+us_births_dataset,T1,Demographic
+saugeenday_dataset,T1,Weather
+london_smart_meters_dataset_without_missing_values,T1,Energy Demand
+london_smart_meters_dataset_without_missing_values,T2,Energy Demand
+london_smart_meters_dataset_without_missing_values,T3,Energy Demand
+traffic_hourly_dataset,T1,Transportation
+traffic_hourly_dataset,T2,Transportation
+traffic_hourly_dataset,T3,Transportation
+traffic_hourly_dataset,T4,Transportation
+traffic_hourly_dataset,T5,Transportation
+electricity_hourly_dataset,T1,Energy Demand
+electricity_hourly_dataset,T2,Energy Demand
+electricity_hourly_dataset,T3,Energy Demand
+pedestrian_counts_dataset,T1,Transportation
+pedestrian_counts_dataset,T2,Transportation
+pedestrian_counts_dataset,T3,Transportation
+pedestrian_counts_dataset,T4,Transportation
+pedestrian_counts_dataset,T5,Transportation
+kdd_cup_2018_dataset_without_missing_values,T1,Other
+australian_electricity_demand_dataset,T1,Energy Demand
+australian_electricity_demand_dataset,T2,Energy Demand
+australian_electricity_demand_dataset,T3,Energy Demand
+oikolab_weather_dataset,T1,Weather
+oikolab_weather_dataset,T2,Weather
+oikolab_weather_dataset,T3,Weather
+oikolab_weather_dataset,T4,Weather
+m4_monthly_dataset,T122,Macro
+m4_monthly_dataset,T145,Macro
+m4_monthly_dataset,T180,Macro
+m4_monthly_dataset,T186,Macro
+m4_monthly_dataset,T17051,Micro
+m4_monthly_dataset,T17088,Micro
+m4_monthly_dataset,T17132,Micro
+m4_monthly_dataset,T17146,Micro
+m4_monthly_dataset,T26710,Demographic
+m4_monthly_dataset,T27138,Industry
+m4_monthly_dataset,T27170,Industry
+m4_monthly_dataset,T27175,Industry
+m4_monthly_dataset,T27186,Industry
+m4_monthly_dataset,T37009,Finance
+m4_monthly_dataset,T37070,Finance
+m4_monthly_dataset,T37238,Finance
+m4_monthly_dataset,T37248,Finance
+m4_monthly_dataset,T47915,Other
+m4_weekly_dataset,T1,Other
+m4_weekly_dataset,T2,Other
+m4_weekly_dataset,T19,Macro
+m4_weekly_dataset,T20,Macro
+m4_weekly_dataset,T21,Macro
+m4_weekly_dataset,T55,Industry
+m4_weekly_dataset,T56,Industry
+m4_weekly_dataset,T60,Finance
+m4_weekly_dataset,T61,Finance
+m4_weekly_dataset,T62,Finance
+m4_weekly_dataset,T224,Demographic
+m4_weekly_dataset,T225,Demographic
+m4_weekly_dataset,T226,Demographic
+m4_weekly_dataset,T227,Demographic
+m4_weekly_dataset,T248,Micro
+m4_weekly_dataset,T249,Micro
+m4_weekly_dataset,T250,Micro
+m4_daily_dataset,T1,Macro
+m4_daily_dataset,T2,Macro
+m4_daily_dataset,T6,Macro
+m4_daily_dataset,T130,Micro
+m4_daily_dataset,T131,Micro
+m4_daily_dataset,T145,Micro
+m4_daily_dataset,T1604,Demographic
+m4_daily_dataset,T1605,Demographic
+m4_daily_dataset,T1606,Demographic
+m4_daily_dataset,T1607,Demographic
+m4_daily_dataset,T1614,Industry
+m4_daily_dataset,T1615,Industry
+m4_daily_dataset,T1634,Industry
+m4_daily_dataset,T1650,Industry
+m4_daily_dataset,T2036,Finance
+m4_daily_dataset,T2037,Finance
+m4_daily_dataset,T2041,Finance
+m4_daily_dataset,T3595,Other
+m4_daily_dataset,T3597,Other
+m4_hourly_dataset,T170,Other
+m4_hourly_dataset,T171,Other
+m4_hourly_dataset,T172,Other

aeon/datasets/__init__.py

@@ -16,7 +16,10 @@
     "load_human_activity_segmentation_datasets",
     # Write functions
     "write_to_ts_file",
+    "write_to_tsf_file",
     "write_to_arff_file",
+    "write_regression_dataset",
+    "write_forecasting_dataset",
     # Single problem loaders
     "load_airline",
     "load_arrow_head",
@@ -58,7 +61,13 @@
     load_from_tsv_file,
     load_regression,
 )
-from aeon.datasets._data_writers import write_to_arff_file, write_to_ts_file
+from aeon.datasets._data_writers import (
+    write_forecasting_dataset,
+    write_regression_dataset,
+    write_to_arff_file,
+    write_to_ts_file,
+    write_to_tsf_file,
+)
 from aeon.datasets._single_problem_loaders import (
     load_acsf1,
     load_airline,

aeon/datasets/_data_loaders.py

@@ -979,7 +979,7 @@ def load_forecasting(name, extract_path=None, return_metadata=False):
     >>> X=load_forecasting("m1_yearly_dataset") # doctest: +SKIP
     """
     # Allow user to have non standard extract path
-    from aeon.datasets.tsf_datasets import tsf_all
+    from aeon.datasets.tsf_datasets import tsf_monash
 
     if extract_path is not None:
         local_module = extract_path
@@ -995,8 +995,8 @@ def load_forecasting(name, extract_path=None, return_metadata=False):
     if name not in get_downloaded_tsf_datasets(extract_path):
         # Dataset is not already present in the datasets directory provided.
         # If it is not there, download and install it.
-        if name in tsf_all.keys():
-            id = tsf_all[name]
+        if name in tsf_monash.keys():
+            id = tsf_monash[name]
             if extract_path is None:
                 local_dirname = "local_data"
             if not os.path.exists(os.path.join(local_module, local_dirname)):