get_dataset_meta_data

get_dataset_meta_data(data_names=None, features=None, url='https://timeseriesclassification.com/aeon-toolkit/metadata.csv')[source]

Retrieve dataset meta data from timeseriesclassification.com.

Metadata includes the following information for each dataset: - Dataset: name of the problem, set the lists in tsc_datasets for valid names. - TrainSize: number of series in the default train set. - TestSize: number of series in the default train set. - Length: length of the series. If the series are not all the same length,

this is set to 0.

  • NumberClasses: number of classes in the problem.

  • Type: nature of the problem, one of PROBLEM_TYPES

  • Channels: number of channels. If univariate, this is 1.

Parameters:
data_nameslist, default=None

List of dataset names to retrieve meta data for. If None, all datasets are retrieved.

featuresString or List, default=None

List of features to retrieve meta data for. Should be a subset of features listed above. Dataset field is always returned.

urlString

default = “https://timeseriesclassification.com/aeon-toolkit/metadata.csv” Location of the csv metadata file.

Returns:
Pandas dataframe containing meta data for each dataset.
Raises:
URLError or HTTPError if the website is not accessible.