UCR

class torchtime.datasets.UCR(name: str, root: str, train: bool = True, download: bool = False, classes: Optional[List[Any]] = None, **kwargs)

UEA & UCR Time Series Classification Repository [Dau et al., 2019].

Parameters:

• root (str) – Root directory of dataset where UCR/<dataset_name> exist.

• name (str) – The name of the dataset to load from the UCR archive.

• train (bool, optional) – If True, creates dataset from TRAIN.ts or TRAIN.arff, otherwise from TEST.ts or TEST.arff.

• download (bool, optional) – If true, downloads the dataset from the internet and puts it in root directory. If dataset is already downloaded, it is not downloaded again.

• classes (list, optional) – A list containing the class labels, where the index in the list indicates the numeric value it should be mapped to.

• transform (callable, optional) – A function/transform that takes in a uni- or multivariate time series and returns a transformed version. E.g, transforms.NaN2Value

• target_transform (callable, optional) – A function/transform that takes in the target and transforms it.

__getitem__(index: int) → Tuple[Any, Any]

Parameters:

index (int) – The index of the sample to return.

Returns:

(series, target) where target is index of the target class.

Return type:

tuple