update hourly prediction lengths
#5
by
kashif
HF staff
- opened
- monash_tsf.py +38 -11
monash_tsf.py
CHANGED
@@ -121,6 +121,7 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
121 |
description="5 time series representing the half hourly electricity demand of 5 states in Australia: Victoria, New South Wales, Queensland, Tasmania and South Australia.",
|
122 |
url="https://zenodo.org/record/4659727",
|
123 |
file_name="australian_electricity_demand_dataset.zip",
|
|
|
124 |
),
|
125 |
MonashTSFBuilderConfig(
|
126 |
name="wind_farms_minutely",
|
@@ -142,6 +143,7 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
142 |
description="Hourly pedestrian counts captured from 66 sensors in Melbourne city starting from May 2009.",
|
143 |
url="https://zenodo.org/record/4656626",
|
144 |
file_name="pedestrian_counts_dataset.zip",
|
|
|
145 |
),
|
146 |
MonashTSFBuilderConfig(
|
147 |
name="vehicle_trips",
|
@@ -191,6 +193,7 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
191 |
description="137 time series representing the solar power production recorded per every 10 minutes in Alabama state in 2006.",
|
192 |
url="https://zenodo.org/record/4656144",
|
193 |
file_name="solar_10_minutes_dataset.zip",
|
|
|
194 |
),
|
195 |
MonashTSFBuilderConfig(
|
196 |
name="solar_weekly",
|
@@ -198,6 +201,7 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
198 |
description="137 time series representing the weekly solar power production in Alabama state in 2006.",
|
199 |
url="https://zenodo.org/record/4656151",
|
200 |
file_name="solar_weekly_dataset.zip",
|
|
|
201 |
),
|
202 |
MonashTSFBuilderConfig(
|
203 |
name="car_parts",
|
@@ -399,10 +403,14 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
399 |
features = datasets.Features(
|
400 |
{
|
401 |
"start": datasets.Value("timestamp[s]"),
|
402 |
-
"target": datasets.Sequence(
|
|
|
|
|
403 |
"feat_static_cat": datasets.Sequence(datasets.Value("uint64")),
|
404 |
# "feat_static_real": datasets.Sequence(datasets.Value("float32")),
|
405 |
-
"feat_dynamic_real": datasets.Sequence(
|
|
|
|
|
406 |
# "feat_dynamic_cat": datasets.Sequence(datasets.Sequence(datasets.Value("uint64"))),
|
407 |
"item_id": datasets.Value("string"),
|
408 |
}
|
@@ -414,7 +422,9 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
414 |
"target": datasets.Sequence(datasets.Value("float32")),
|
415 |
"feat_static_cat": datasets.Sequence(datasets.Value("uint64")),
|
416 |
# "feat_static_real": datasets.Sequence(datasets.Value("float32")),
|
417 |
-
"feat_dynamic_real": datasets.Sequence(
|
|
|
|
|
418 |
# "feat_dynamic_cat": datasets.Sequence(datasets.Sequence(datasets.Value("uint64"))),
|
419 |
"item_id": datasets.Value("string"),
|
420 |
}
|
@@ -474,7 +484,7 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
474 |
prediction_length_map = {
|
475 |
"S": 60,
|
476 |
"T": 60,
|
477 |
-
"H":
|
478 |
"D": 30,
|
479 |
"W": 8,
|
480 |
"M": 12,
|
@@ -496,13 +506,17 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
496 |
start = ts.start_timestamp[0]
|
497 |
|
498 |
if self.config.target_fields is not None:
|
499 |
-
target_fields = ts[
|
|
|
|
|
500 |
else:
|
501 |
target_fields = self.config.data_column.unique()
|
502 |
|
503 |
if self.config.feat_dynamic_real_fields is not None:
|
504 |
feat_dynamic_real_fields = ts[
|
505 |
-
ts[self.config.data_column].isin(
|
|
|
|
|
506 |
]
|
507 |
feat_dynamic_real = np.vstack(feat_dynamic_real_fields.target)
|
508 |
else:
|
@@ -513,7 +527,10 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
513 |
feat_static_cat = [cat]
|
514 |
|
515 |
if split in ["train", "val"]:
|
516 |
-
offset =
|
|
|
|
|
|
|
517 |
target = target[..., :-offset]
|
518 |
if self.config.feat_dynamic_real_fields is not None:
|
519 |
feat_dynamic_real = feat_dynamic_real[..., :-offset]
|
@@ -527,18 +544,25 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
527 |
}
|
528 |
else:
|
529 |
if self.config.target_fields is not None:
|
530 |
-
target_fields = loaded_data[
|
|
|
|
|
531 |
else:
|
532 |
target_fields = loaded_data
|
533 |
if self.config.feat_dynamic_real_fields is not None:
|
534 |
feat_dynamic_real_fields = loaded_data[
|
535 |
-
loaded_data[self.config.data_column].isin(
|
|
|
|
|
536 |
]
|
537 |
else:
|
538 |
feat_dynamic_real_fields = None
|
539 |
|
540 |
for cat, ts in target_fields.iterrows():
|
541 |
-
start = ts.get(
|
|
|
|
|
|
|
542 |
target = ts.target
|
543 |
if feat_dynamic_real_fields is not None:
|
544 |
feat_dynamic_real = np.vstack(feat_dynamic_real_fields.target)
|
@@ -552,7 +576,10 @@ class MonashTSF(datasets.GeneratorBasedBuilder):
|
|
552 |
item_id = ts.series_name
|
553 |
|
554 |
if split in ["train", "val"]:
|
555 |
-
offset =
|
|
|
|
|
|
|
556 |
target = target[..., :-offset]
|
557 |
if feat_dynamic_real is not None:
|
558 |
feat_dynamic_real = feat_dynamic_real[..., :-offset]
|
|
|
121 |
description="5 time series representing the half hourly electricity demand of 5 states in Australia: Victoria, New South Wales, Queensland, Tasmania and South Australia.",
|
122 |
url="https://zenodo.org/record/4659727",
|
123 |
file_name="australian_electricity_demand_dataset.zip",
|
124 |
+
prediction_length=336,
|
125 |
),
|
126 |
MonashTSFBuilderConfig(
|
127 |
name="wind_farms_minutely",
|
|
|
143 |
description="Hourly pedestrian counts captured from 66 sensors in Melbourne city starting from May 2009.",
|
144 |
url="https://zenodo.org/record/4656626",
|
145 |
file_name="pedestrian_counts_dataset.zip",
|
146 |
+
prediction_length=24,
|
147 |
),
|
148 |
MonashTSFBuilderConfig(
|
149 |
name="vehicle_trips",
|
|
|
193 |
description="137 time series representing the solar power production recorded per every 10 minutes in Alabama state in 2006.",
|
194 |
url="https://zenodo.org/record/4656144",
|
195 |
file_name="solar_10_minutes_dataset.zip",
|
196 |
+
prediction_length=1008,
|
197 |
),
|
198 |
MonashTSFBuilderConfig(
|
199 |
name="solar_weekly",
|
|
|
201 |
description="137 time series representing the weekly solar power production in Alabama state in 2006.",
|
202 |
url="https://zenodo.org/record/4656151",
|
203 |
file_name="solar_weekly_dataset.zip",
|
204 |
+
prediction_length=5,
|
205 |
),
|
206 |
MonashTSFBuilderConfig(
|
207 |
name="car_parts",
|
|
|
403 |
features = datasets.Features(
|
404 |
{
|
405 |
"start": datasets.Value("timestamp[s]"),
|
406 |
+
"target": datasets.Sequence(
|
407 |
+
datasets.Sequence(datasets.Value("float32"))
|
408 |
+
),
|
409 |
"feat_static_cat": datasets.Sequence(datasets.Value("uint64")),
|
410 |
# "feat_static_real": datasets.Sequence(datasets.Value("float32")),
|
411 |
+
"feat_dynamic_real": datasets.Sequence(
|
412 |
+
datasets.Sequence(datasets.Value("float32"))
|
413 |
+
),
|
414 |
# "feat_dynamic_cat": datasets.Sequence(datasets.Sequence(datasets.Value("uint64"))),
|
415 |
"item_id": datasets.Value("string"),
|
416 |
}
|
|
|
422 |
"target": datasets.Sequence(datasets.Value("float32")),
|
423 |
"feat_static_cat": datasets.Sequence(datasets.Value("uint64")),
|
424 |
# "feat_static_real": datasets.Sequence(datasets.Value("float32")),
|
425 |
+
"feat_dynamic_real": datasets.Sequence(
|
426 |
+
datasets.Sequence(datasets.Value("float32"))
|
427 |
+
),
|
428 |
# "feat_dynamic_cat": datasets.Sequence(datasets.Sequence(datasets.Value("uint64"))),
|
429 |
"item_id": datasets.Value("string"),
|
430 |
}
|
|
|
484 |
prediction_length_map = {
|
485 |
"S": 60,
|
486 |
"T": 60,
|
487 |
+
"H": 168,
|
488 |
"D": 30,
|
489 |
"W": 8,
|
490 |
"M": 12,
|
|
|
506 |
start = ts.start_timestamp[0]
|
507 |
|
508 |
if self.config.target_fields is not None:
|
509 |
+
target_fields = ts[
|
510 |
+
ts[self.config.data_column].isin(self.config.target_fields)
|
511 |
+
]
|
512 |
else:
|
513 |
target_fields = self.config.data_column.unique()
|
514 |
|
515 |
if self.config.feat_dynamic_real_fields is not None:
|
516 |
feat_dynamic_real_fields = ts[
|
517 |
+
ts[self.config.data_column].isin(
|
518 |
+
self.config.feat_dynamic_real_fields
|
519 |
+
)
|
520 |
]
|
521 |
feat_dynamic_real = np.vstack(feat_dynamic_real_fields.target)
|
522 |
else:
|
|
|
527 |
feat_static_cat = [cat]
|
528 |
|
529 |
if split in ["train", "val"]:
|
530 |
+
offset = (
|
531 |
+
forecast_horizon * self.config.rolling_evaluations
|
532 |
+
+ forecast_horizon * (split == "train")
|
533 |
+
)
|
534 |
target = target[..., :-offset]
|
535 |
if self.config.feat_dynamic_real_fields is not None:
|
536 |
feat_dynamic_real = feat_dynamic_real[..., :-offset]
|
|
|
544 |
}
|
545 |
else:
|
546 |
if self.config.target_fields is not None:
|
547 |
+
target_fields = loaded_data[
|
548 |
+
loaded_data[self.config.data_column].isin(self.config.target_fields)
|
549 |
+
]
|
550 |
else:
|
551 |
target_fields = loaded_data
|
552 |
if self.config.feat_dynamic_real_fields is not None:
|
553 |
feat_dynamic_real_fields = loaded_data[
|
554 |
+
loaded_data[self.config.data_column].isin(
|
555 |
+
self.config.feat_dynamic_real_fields
|
556 |
+
)
|
557 |
]
|
558 |
else:
|
559 |
feat_dynamic_real_fields = None
|
560 |
|
561 |
for cat, ts in target_fields.iterrows():
|
562 |
+
start = ts.get(
|
563 |
+
"start_timestamp",
|
564 |
+
datetime.strptime("1900-01-01 00-00-00", "%Y-%m-%d %H-%M-%S"),
|
565 |
+
)
|
566 |
target = ts.target
|
567 |
if feat_dynamic_real_fields is not None:
|
568 |
feat_dynamic_real = np.vstack(feat_dynamic_real_fields.target)
|
|
|
576 |
item_id = ts.series_name
|
577 |
|
578 |
if split in ["train", "val"]:
|
579 |
+
offset = (
|
580 |
+
forecast_horizon * self.config.rolling_evaluations
|
581 |
+
+ forecast_horizon * (split == "train")
|
582 |
+
)
|
583 |
target = target[..., :-offset]
|
584 |
if feat_dynamic_real is not None:
|
585 |
feat_dynamic_real = feat_dynamic_real[..., :-offset]
|