orion.primitives.azure_anomaly_detector.split_sequence

orion.primitives.azure_anomaly_detector.split_sequence(X, index, target_column, sequence_size, overlap_size)[source]

Split sequences of time series data.

The function creates a list of input sequences by splitting the input sequence into partitions with a specified size and pads it with values from previous sequence according to the overlap size.

Parameters
  • X (ndarray) – N-dimensional value sequence to iterate over.

  • index (ndarray) – N-dimensional index sequence to iterate over.

  • target_column (int) – Indicating which column of X is the target.

  • sequence_size (int) – Length of the input sequences.

  • overlap_size (int) – Length of the values from previous window.

Returns

  • List of sliced value as ndarray.

  • List of sliced index as ndarray.

Return type

tuple