BiTCN - Nixtla

Bidirectional Temporal Convolutional Network (BiTCN) is a forecasting architecture based on two temporal convolutional networks (TCNs). The first network (‘forward’) encodes future covariates of the time series, whereas the second network (‘backward’) encodes past observations and covariates. This method allows to preserve the temporal information of sequence data, and is computationally more efficient than common RNN methods (LSTM, GRU, …). As compared to Transformer-based methods, BiTCN has a lower space complexity, i.e. it requires orders of magnitude less parameters. This model may be a good choice if you seek a small model (small amount of trainable parameters) with few hyperparameters to tune (only 2). References

Figure 1. Visualization of a stack of dilated causal convolutional layers.

1. BiTCN

`BiTCN`

BiTCN(
    h,
    input_size,
    hidden_size=16,
    dropout=0.5,
    futr_exog_list=None,
    hist_exog_list=None,
    stat_exog_list=None,
    exclude_insample_y=False,
    loss=MAE(),
    valid_loss=None,
    max_steps=1000,
    learning_rate=0.001,
    num_lr_decays=-1,
    early_stop_patience_steps=-1,
    val_check_steps=100,
    batch_size=32,
    valid_batch_size=None,
    windows_batch_size=1024,
    inference_windows_batch_size=1024,
    start_padding_enabled=False,
    training_data_availability_threshold=0.0,
    step_size=1,
    scaler_type="identity",
    random_seed=1,
    drop_last_loader=False,
    alias=None,
    optimizer=None,
    optimizer_kwargs=None,
    lr_scheduler=None,
    lr_scheduler_kwargs=None,
    dataloader_kwargs=None,
    **trainer_kwargs
)

Bases: BaseModel BiTCN Bidirectional Temporal Convolutional Network (BiTCN) is a forecasting architecture based on two temporal convolutional networks (TCNs). The first network (‘forward’) encodes future covariates of the time series, whereas the second network (‘backward’) encodes past observations and covariates. This is a univariate model. Parameters:

Name	Type	Description	Default
`h`	`int`	forecast horizon.	required
`input_size`	`int`	considered autorregresive inputs (lags), y=[1,2,3,4] input_size=2 -> lags=[1,2].	required
`hidden_size`	`int`	units for the TCN’s hidden state size. Default: 16.	`16`
`dropout`	`float`	dropout rate used for the dropout layers throughout the architecture. Default: 0.1.	`0.5`
`futr_exog_list`	`list`	future exogenous columns.	`None`
`hist_exog_list`	`list`	historic exogenous columns.	`None`
`stat_exog_list`	`list`	static exogenous columns.	`None`
`exclude_insample_y`	`bool`	the model skips the autoregressive features y[t-input_size:t] if True. Default: False.	`False`
`loss`	`Module`	PyTorch module, instantiated train loss class from losses collection.	`MAE()`
`valid_loss`	`Module`	PyTorch module, instantiated valid loss class from losses collection.	`None`
`max_steps`	`int`	maximum number of training steps. Default: 1000.	`1000`
`learning_rate`	`float`	Learning rate between (0, 1). Default: 1e-3.	`0.001`
`num_lr_decays`	`int`	Number of learning rate decays, evenly distributed across max_steps. Default: -1.	`-1`
`early_stop_patience_steps`	`int`	Number of validation iterations before early stopping. Default: -1.	`-1`
`val_check_steps`	`int`	Number of training steps between every validation loss check. Default: 100.	`100`
`batch_size`	`int`	number of different series in each batch. Default: 32.	`32`
`valid_batch_size`	`int`	number of different series in each validation and test batch, if None uses batch_size. Default: None.	`None`
`windows_batch_size`	`int`	number of windows to sample in each training batch, default uses all. Default: 1024.	`1024`
`inference_windows_batch_size`	`int`	number of windows to sample in each inference batch, -1 uses all. Default: 1024.	`1024`
`start_padding_enabled`	`bool`	if True, the model will pad the time series with zeros at the beginning, by input size. Default: False.	`False`
`training_data_availability_threshold`	`Union[float, List[float]]`	minimum fraction of valid data points required for training windows. Single float applies to both insample and outsample; list of two floats specifies [insample_fraction, outsample_fraction]. Default 0.0 allows windows with only 1 valid data point (current behavior). Default: 0.0.	`0.0`
`step_size`	`int`	step size between each window of temporal data. Default: 1.	`1`
`scaler_type`	`str`	type of scaler for temporal inputs normalization see temporal scalers. Default: ‘identity’.	`‘identity’`
`random_seed`	`int`	random_seed for pytorch initializer and numpy generators. Default: 1.	`1`
`drop_last_loader`	`bool`	if True `TimeSeriesDataLoader` drops last non-full batch. Default: False.	`False`
`alias`	`str`	optional, Custom name of the model. Default: None.	`None`
`optimizer`	`Subclass of ‘torch.optim.Optimizer’`	optional, user specified optimizer instead of the default choice (Adam).	`None`
`optimizer_kwargs`	`dict`	optional, list of parameters used by the user specified `optimizer`.	`None`
`lr_scheduler`	`Subclass of ‘torch.optim.lr_scheduler.LRScheduler’`	optional, user specified lr_scheduler instead of the default choice (StepLR).	`None`
`lr_scheduler_kwargs`	`dict`	optional, list of parameters used by the user specified `lr_scheduler`.	`None`
`dataloader_kwargs`	`dict`	optional, list of parameters passed into the PyTorch Lightning dataloader by the `TimeSeriesDataLoader`.	`None`
`**trainer_kwargs`	`int`	keyword trainer arguments inherited from PyTorch Lighning’s trainer.

`BiTCN.fit`

fit(
    dataset, val_size=0, test_size=0, random_seed=None, distributed_config=None
)

Fit. The fit method, optimizes the neural network’s weights using the initialization parameters (learning_rate, windows_batch_size, …) and the loss function as defined during the initialization. Within fit we use a PyTorch Lightning Trainer that inherits the initialization’s self.trainer_kwargs, to customize its inputs, see PL’s trainer arguments. The method is designed to be compatible with SKLearn-like classes and in particular to be compatible with the StatsForecast library. By default the model is not saving training checkpoints to protect disk memory, to get them change enable_checkpointing=True in __init__. Parameters:

Name	Type	Description	Default
`dataset`	`TimeSeriesDataset`	NeuralForecast’s `TimeSeriesDataset`, see documentation.	required
`val_size`	`int`	Validation size for temporal cross-validation.	`0`
`random_seed`	`int`	Random seed for pytorch initializer and numpy generators, overwrites model.init’s.	`None`
`test_size`	`int`	Test size for temporal cross-validation.	`0`

Returns:

Type	Description
None

`BiTCN.predict`

predict(
    dataset,
    test_size=None,
    step_size=1,
    random_seed=None,
    quantiles=None,
    h=None,
    explainer_config=None,
    **data_module_kwargs
)

Predict. Neural network prediction with PL’s Trainer execution of predict_step. Parameters:

Name	Type	Description	Default
`dataset`	`TimeSeriesDataset`	NeuralForecast’s `TimeSeriesDataset`, see documentation.	required
`test_size`	`int`	Test size for temporal cross-validation.	`None`
`step_size`	`int`	Step size between each window.	`1`
`random_seed`	`int`	Random seed for pytorch initializer and numpy generators, overwrites model.init’s.	`None`
`quantiles`	`list`	Target quantiles to predict.	`None`
`h`	`int`	Prediction horizon, if None, uses the model’s fitted horizon. Defaults to None.	`None`
`explainer_config`	`dict`	configuration for explanations.	`None`
`**data_module_kwargs`	`dict`	PL’s TimeSeriesDataModule args, see documentation.

Returns:

Type	Description
None

Usage Example

import pandas as pd
import matplotlib.pyplot as plt

from neuralforecast import NeuralForecast
from neuralforecast.losses.pytorch import GMM
from neuralforecast.models import BiTCN
from neuralforecast.utils import AirPassengersPanel, AirPassengersStatic

Y_train_df = AirPassengersPanel[AirPassengersPanel.ds<AirPassengersPanel['ds'].values[-12]] # 132 train
Y_test_df = AirPassengersPanel[AirPassengersPanel.ds>=AirPassengersPanel['ds'].values[-12]].reset_index(drop=True) # 12 test

fcst = NeuralForecast(
    models=[
            BiTCN(h=12,
                input_size=24,
                loss=GMM(n_components=7, level=[80,90]),
                max_steps=100,
                scaler_type='standard',
                futr_exog_list=['y_[lag12]'],
                hist_exog_list=None,
                stat_exog_list=['airline1'],
                windows_batch_size=2048,
                val_check_steps=10,
                early_stop_patience_steps=-1,
                ),     
    ],
    freq='ME'
)
fcst.fit(df=Y_train_df, static_df=AirPassengersStatic)
forecasts = fcst.predict(futr_df=Y_test_df)

# Plot quantile predictions
Y_hat_df = forecasts.reset_index(drop=False).drop(columns=['unique_id','ds'])
plot_df = pd.concat([Y_test_df, Y_hat_df], axis=1)
plot_df = pd.concat([Y_train_df, plot_df])

plot_df = plot_df[plot_df.unique_id=='Airline1'].drop('unique_id', axis=1)
plt.plot(plot_df['ds'], plot_df['y'], c='black', label='True')
plt.plot(plot_df['ds'], plot_df['BiTCN-median'], c='blue', label='median')
plt.fill_between(x=plot_df['ds'][-12:], 
                 y1=plot_df['BiTCN-lo-90'][-12:].values,
                 y2=plot_df['BiTCN-hi-90'][-12:].values,
                 alpha=0.4, label='level 90')
plt.legend()
plt.grid()

2. Auxilary functions

`TCNCell`

TCNCell(
    in_channels,
    out_channels,
    kernel_size,
    padding,
    dilation,
    mode,
    groups,
    dropout,
)

Bases: Module Temporal Convolutional Network Cell, consisting of CustomConv1D modules.

`CustomConv1d`

CustomConv1d(
    in_channels,
    out_channels,
    kernel_size,
    padding=0,
    dilation=1,
    mode="backward",
    groups=1,
)

Bases: Module Forward- and backward looking Conv1D

​1. BiTCN

​BiTCN

​BiTCN.fit

​BiTCN.predict

​Usage Example

​2. Auxilary functions

​TCNCell

​CustomConv1d