TranAD#

class oats.models.reconstruction.tranad.TranADModel(window: int = 100, use_gpu: bool = False, val_split: float = 0.2)[source]#

Bases: Model

TranAD Model

Tuli, Shreshth and Casale, Giuliano and Jennings, Nicholas R “TranAD: Deep Transformer Networks for Anomaly Detection in Multivariate Time Series Data”

Implementation from reference: https://github.com/imperial-qore/TranAD

Parameters:

  • window (int, optional) – rolling window size to feed into the predictor. Defaults to 10.

  • use_gpu (bool, optional) – whether to use GPU. Defaults to False.

  • val_split (float, optional) – proportion of data points reserved for validation. Defaults to 0.2.

fit(train_data, epochs=5)[source]#
get_scores(test_data)[source]#