How can I create training, validation, and test sets like "tourism_monthly"
I have a continuous dataset of about 120,000 in length in gluonts ListDataset format. How can I create training, validation, and test sets, where the number of validation sequences is the same as the training sequences, and each validation sequence is the same as the training sequence in the front, with an extra prediction_length at the tail, just like the dataset ("monash_tsf", "tourism_monthly")?
example:https://huggingface.co/blog/time-series-transformers,
len(train_example["target"]) + prediction_length is equal to len( validation_example["target"]
@byjwangyu so to understand you have a dataset with a single time series?
@byjwangyu indeed have a look at: https://github.com/huggingface/notebooks/blob/main/examples/time_series_datasets.ipynb
Hello. Is there any notebook available for multivariate time series dataset preparation?
I have a dataset with many features and only one target feature.
@kashif
Brother could you help me in this regard?