def comparisonExperiments():
data_dir = sys.argv[1]
res_dir = sys.argv[2]
complete_classifiers = [
"ShapeletForestClassifier",
]
small_datasets = [
"Beef",
"Car",
"Coffee",
"CricketX",
"CricketY",
"CricketZ",
"DiatomSizeReduction",
"Fish",
"GunPoint",
"ItalyPowerDemand",
"MoteStrain",
"OliveOil",
"Plane",
"SonyAIBORobotSurface1",
"SonyAIBORobotSurface2",
"SyntheticControl",
"Trace",
"TwoLeadECG",
]
small_datasets = [
"Beef",
"Coffee",
]
num_folds = 2
import sktime.contrib.experiments as exp
for f in range(num_folds):
for d in small_datasets:
for c in complete_classifiers:
print(c, d, f)
try:
exp.run_experiment(data_dir, res_dir, c, d, f)
except:
print('\n\n FAILED: ', sys.exc_info()[0], '\n\n')
def comparisonExperiments():
data_dir = '/home/carl/Downloads/Univariate2018_ts/Univariate_ts/' # sys.argv[1]
res_dir = '/home/carl/temp/' # sys.argv[2]
complete_classifiers = [
"catch22ForestClassifier",
]
small_datasets = [
"Beef",
"Car",
"Coffee",
"CricketX",
"CricketY",
"CricketZ",
"DiatomSizeReduction",
"Fish",
"GunPoint",
"ItalyPowerDemand",
"MoteStrain",
"OliveOil",
"Plane",
"SonyAIBORobotSurface1",
"SonyAIBORobotSurface2",
"SyntheticControl",
"Trace",
"TwoLeadECG",
]
small_datasets = [
"Beef",
"Coffee",
]
num_folds = 2
import sktime.contrib.experiments as exp
for f in range(num_folds):
for d in small_datasets:
for c in complete_classifiers:
print(c, d, f)
# try:
exp.run_experiment(data_dir, res_dir, c, d, f)
def tsf_benchmarking():
for i in range(len(benchmark_datasets)):
dataset = benchmark_datasets[i]
print(str(i) + " problem = " + dataset)
tsf = ib.TimeSeriesForest(n_estimators=100)
exp.run_experiment(overwrite=False,
problem_path=data_dir,
results_path=results_dir,
cls_name="PythonTSF",
classifier=tsf,
dataset=dataset,
train_file=False)
steps = [
('segment', RandomIntervalSegmenter(n_intervals='sqrt')),
('transform',
FeatureUnion([('mean',
RowTransformer(
FunctionTransformer(func=np.mean,
validate=False))),
('std',
RowTransformer(
FunctionTransformer(func=np.std,
validate=False))),
('slope',
RowTransformer(
FunctionTransformer(func=time_series_slope,
validate=False)))])),
('clf', DecisionTreeClassifier())
]
base_estimator = Pipeline(steps)
tsf = TimeSeriesForestClassifier(estimator=base_estimator,
n_estimators=100)
exp.run_experiment(overwrite=False,
problem_path=data_dir,
results_path=results_dir,
cls_name="PythonTSFComposite",
classifier=tsf,
dataset=dataset,
train_file=False)
def elastic_distance_benchmarking():
for i in range(int(len(distance_test))):
dataset = distance_test[i]
print(
str(i) + " problem = " + dataset + " writing to " + results_dir +
"/DTW/")
dtw = dist.KNeighborsTimeSeriesClassifier(metric="dtw")
exp.run_experiment(overwrite=False,
problem_path=data_dir,
results_path=results_dir + "/DTW/",
cls_name="PythonDTW",
classifier=dtw,
dataset=dataset,
train_file=False)
twe = dist.KNeighborsTimeSeriesClassifier(metric="dtw")
exp.run_experiment(overwrite=False,
problem_path=data_dir,
results_path=results_dir + "/DTW/",
cls_name="PythonTWE",
classifier=twe,
dataset=dataset,
train_file=False)
def dlExperiment(data_dir, res_dir, classifier_name, dset, fold, classifier=None):
if classifier is None:
classifier = setNetwork(classifier_name, fold)
exp.run_experiment(data_dir, res_dir, classifier_name, dset, classifier=classifier, resampleID=fold)
if __name__ == "__main__":
"""
Example simple usage, with arguments input via script or hard coded for testing
"""
print(" Local Run")
results_dir = "C:/Temp/sktime-dl/"
classifier = "resnet"
resample = 0
# for i in range(0, len(univariate_datasets)):
# dataset = univariate_datasets[i]
# # print(i)
# # print(" problem = "+dataset)
problem = "ItalyPowerDemand"
print("Loading ", problem)
trX, trY = load_UCR_UEA_dataset(problem, split="train", return_X_y=True)
teX, teY = load_UCR_UEA_dataset(problem, split="test", return_X_y=True)
tf = False
run_experiment(
overwrite=True,
trainX=trX,
trainY=trY,
testX=trX,
testY=trY,
results_path=results_dir,
cls_name=classifier,
dataset=problem,
resampleID=resample,
train_file=tf,
)