def _load_dataset(self, dataset_name):
if dataset_name == "bc-data":
self._dataset = bc.load()
elif dataset_name == "total-load":
self._dataset = ul.total_experiment_load()[1]['Load']
else:
raise RuntimeError("Invalid dataset name: %s" % dataset_name)
def _load_dataset(self, dataset_name):
if dataset_name == "bc-data":
self._dataset = bc.load()
elif dataset_name == "total-load":
self._dataset = ul.total_experiment_load()[1]['Load']
else:
raise RuntimeError("Invalid dataset name: %s" % dataset_name)
def _prepare_dataset(self):
"""Return datasets for the experiment periods containing the total load
for all users in the experiment."""
print "Using total load rather than per-user load."
temps = read_temperatures()
loads = pd.concat(ul.total_experiment_load())
return [self._join_temp_and_load(temps, loads, period)
for period in ul.experiment_periods()]
def _prepare_dataset(self):
"""Return datasets for the experiment periods containing the total load
for all users in the experiment."""
loads = pd.concat(ul.total_experiment_load())
return [
ul.add_temperatures(loads, period)
for period in ul.experiment_periods()
]
def _prepare_dataset(self):
"""Return datasets for the experiment periods containing the total load
for all users in the experiment."""
print "Using total load rather than per-user load."
temps = read_temperatures()
loads = pd.concat(ul.total_experiment_load())
return [
self._join_temp_and_load(temps, loads, period)
for period in ul.experiment_periods()
]
def _get_smoother():
# Set slow_smoother to True in order to see the actual time consumed by the
# B-spline smoothing operation. If set to False, will use the default
# smoother where the roughness matrices are cached.
slow_smoother = True
if slow_smoother:
print "Using slow, analytic, non-caching smoother."
return cln.BSplineAnalyticSmoother
else:
print "Using not quite so slow, caching smoother."
return cln.BSplineSmoother
# Load a dataset containing power load history. This set is divided into
# training and test data, we only keep the traning part for now.
dataset, _ = ul.total_experiment_load()
# Set parameters for the B-spline smoother/cleanser
smoothness = 10
zscore = 0.5
# Try smoothing/cleansing different time series lengths
for hindsight_days in [1]:
# Select data
num_hours = 24 * hindsight_days
data = dataset["Load"][-num_hours:].copy()
# Some output and rough timing
print "Cleansing %d hours of data with smoothness %.2f, z-score %.2f..." % (num_hours, smoothness, zscore)
sys.stdout.flush()
start_time = time.time()
# This is the part that takes time
smoother = _get_smoother()(data, smoothness)
def _get_smoother():
# Set slow_smoother to True in order to see the actual time consumed by the
# B-spline smoothing operation. If set to False, will use the default
# smoother where the roughness matrices are cached.
slow_smoother = True
if slow_smoother:
print "Using slow, analytic, non-caching smoother."
return cln.BSplineAnalyticSmoother
else:
print "Using not quite so slow, caching smoother."
return cln.BSplineSmoother
# Load a dataset containing power load history. This set is divided into
# training and test data, we only keep the traning part for now.
dataset, _ = ul.total_experiment_load()
# Set parameters for the B-spline smoother/cleanser
smoothness = 10
zscore = 0.5
# Try smoothing/cleansing different time series lengths
for hindsight_days in [1]:
# Select data
num_hours = 24 * hindsight_days
data = dataset["Load"][-num_hours:].copy()
# Some output and rough timing
print "Cleansing %d hours of data with smoothness %.2f, z-score %.2f..." % \
(num_hours, smoothness, zscore)
sys.stdout.flush()
start_time = time.time()
# This is the part that takes time
def _prepare_dataset(self):
"""Return datasets for the experiment periods containing the total load
for all users in the experiment."""
loads = pd.concat(ul.total_experiment_load())
return [ul.add_temperatures(loads, period)
for period in ul.experiment_periods()]
def setUpClass(cls):
cls.data = ul.add_temperatures(ul.total_experiment_load()[0],
ul.experiment_periods()[0])
