citylearn.end_use_load_profiles.lstm_model.preprocessing module

citylearn.end_use_load_profiles.lstm_model.preprocessing.dataset_dataloader(x: ndarray, y: ndarray, batch_size: int, shuffle: bool = None, drop_last: bool = None) → Tuple[TensorDataset, DataLoader]

citylearn.end_use_load_profiles.lstm_model.preprocessing.preprocess_df(config: Mapping[str, Any], df: DataFrame, train_references: List[int] = None, validation_references: List[int] = None, test_references: List[int] = None) → Mapping[str, Any]

citylearn.end_use_load_profiles.lstm_model.preprocessing.sliding_windows(data: ndarray, seq_length: int, output_len: int)

Create sliding window input/output arrays.

The target variable is expected to be the last column in data.

Parameters:

• data – Input array.

• seq_length – Number of lookback time steps.

• output_len – Number of time steps ahead to predict.

Returns:

Tuple (x, y) where x has shape [num_steps - seq_length, seq_length, num_input_variables] and y has shape [num_steps - seq_length, output_len].