PAR.sample(num_sequences=None, context=None, sequence_length=None)

Sample new sequences.

  • num_sequences (int) – Number of sequences to sample. If context is passed, this is ignored. If not given, the same number of sequences as in the original timeseries_data is sampled.

  • context (pandas.DataFrame) – Context values to use when generating the sequences. If not passed, the context values will be sampled using the specified tabular model.

  • sequence_length (int) – If passed, sample sequences of this length. If not given, the sequence length will be sampled from the model.


Table containing the sampled sequences in the same format as that he training data had.

Return type