The TimesNet univariate model tackles the challenge of modeling multiple intraperiod and interperiod temporal variations.

The architecture has the following distinctive features: - An embedding layer that maps the input sequence into a latent space. - Transformation of 1D time seires into 2D tensors, based on periods found by FFT. - A convolutional Inception block that captures temporal variations at different scales and between periods.

References
- Haixu Wu and Tengge Hu and Yong Liu and Hang Zhou and Jianmin Wang and Mingsheng Long. TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis - Based on the implementation in https://github.com/thuml/Time-Series-Library (license: https://github.com/thuml/Time-Series-Library/blob/main/LICENSE)

​
1. Auxiliary Functions

source

​
Inception_Block_V1

 Inception_Block_V1 (in_channels, out_channels, num_kernels=6,
                     init_weight=True)

Inception_Block_V1

source

​
TimesBlock

 TimesBlock (input_size, h, k, hidden_size, conv_hidden_size, num_kernels)

TimesBlock

source

​
FFT_for_Period

 FFT_for_Period (x, k=2)

​
2. TimesNet

source

​
TimesNet

 TimesNet (h:int, input_size:int, stat_exog_list=None,
           hist_exog_list=None, futr_exog_list=None,
           exclude_insample_y=False, hidden_size:int=64,
           dropout:float=0.1, conv_hidden_size:int=64, top_k:int=5,
           num_kernels:int=6, encoder_layers:int=2, loss=MAE(),
           valid_loss=None, max_steps:int=1000,
           learning_rate:float=0.0001, num_lr_decays:int=-1,
           early_stop_patience_steps:int=-1, val_check_steps:int=100,
           batch_size:int=32, valid_batch_size:Optional[int]=None,
           windows_batch_size=64, inference_windows_batch_size=256,
           start_padding_enabled=False, step_size:int=1,
           scaler_type:str='standard', random_seed:int=1,
           drop_last_loader:bool=False, alias:Optional[str]=None,
           optimizer=None, optimizer_kwargs=None, lr_scheduler=None,
           lr_scheduler_kwargs=None, dataloader_kwargs=None,
           **trainer_kwargs)

*TimesNet

The TimesNet univariate model tackles the challenge of modeling multiple intraperiod and interperiod temporal variations.

Parameters
h : int, Forecast horizon.
input_size : int, Length of input window (lags).
stat_exog_list : list of str, optional (default=None), Static exogenous columns.
hist_exog_list : list of str, optional (default=None), Historic exogenous columns.
futr_exog_list : list of str, optional (default=None), Future exogenous columns.
exclude_insample_y : bool (default=False), The model skips the autoregressive features y[t-input_size:t] if True.
hidden_size : int (default=64), Size of embedding for embedding and encoders.
dropout : float between [0, 1) (default=0.1), Dropout for embeddings.
conv_hidden_size: int (default=64), Channels of the Inception block.
top_k: int (default=5), Number of periods.
num_kernels: int (default=6), Number of kernels for the Inception block.
encoder_layers : int, (default=2), Number of encoder layers.
loss: PyTorch module (default=MAE()), Instantiated train loss class from losses collection. valid_loss: PyTorch module (default=None, uses loss), Instantiated validation loss class from losses collection.
max_steps: int (default=1000), Maximum number of training steps.
learning_rate : float (default=1e-4), Learning rate.
num_lr_decays: int (default=-1), Number of learning rate decays, evenly distributed across max_steps. If -1, no learning rate decay is performed.
early_stop_patience_steps : int (default=-1), Number of validation iterations before early stopping. If -1, no early stopping is performed.
val_check_steps : int (default=100), Number of training steps between every validation loss check.
batch_size : int (default=32), Number of different series in each batch.
valid_batch_size : int (default=None), Number of different series in each validation and test batch, if None uses batch_size.
windows_batch_size : int (default=64), Number of windows to sample in each training batch.
inference_windows_batch_size : int (default=256), Number of windows to sample in each inference batch.
start_padding_enabled : bool (default=False), If True, the model will pad the time series with zeros at the beginning by input size.
step_size : int (default=1), Step size between each window of temporal data.
scaler_type : str (default=‘standard’), Type of scaler for temporal inputs normalization see temporal scalers.
random_seed : int (default=1), Random_seed for pytorch initializer and numpy generators.
drop_last_loader : bool (default=False), If True TimeSeriesDataLoader drops last non-full batch.
alias : str, optional (default=None), Custom name of the model.
optimizer: Subclass of ‘torch.optim.Optimizer’, optional (default=None), User specified optimizer instead of the default choice (Adam).
optimizer_kwargs: dict, optional (defualt=None), List of parameters used by the user specified optimizer.
lr_scheduler: Subclass of ‘torch.optim.lr_scheduler.LRScheduler’, optional, user specified lr_scheduler instead of the default choice (StepLR).
lr_scheduler_kwargs: dict, optional, list of parameters used by the user specified lr_scheduler.

dataloader_kwargs: dict, optional (default=None), List of parameters passed into the PyTorch Lightning dataloader by the TimeSeriesDataLoader.
**trainer_kwargs: Keyword trainer arguments inherited from PyTorch Lighning’s trainer*

​
TimesNet.fit

 TimesNet.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:
dataset: NeuralForecast’s TimeSeriesDataset, see documentation.
val_size: int, validation size for temporal cross-validation.
random_seed: int=None, random_seed for pytorch initializer and numpy generators, overwrites model.__init__’s.
test_size: int, test size for temporal cross-validation.
*

​
TimesNet.predict

 TimesNet.predict (dataset, test_size=None, step_size=1, random_seed=None,
                   quantiles=None, **data_module_kwargs)

*Predict.

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

Parameters:
dataset: NeuralForecast’s TimeSeriesDataset, see documentation.
test_size: int=None, test size for temporal cross-validation.
step_size: int=1, Step size between each window.
random_seed: int=None, random_seed for pytorch initializer and numpy generators, overwrites model.__init__’s.
quantiles: list of floats, optional (default=None), target quantiles to predict.
**data_module_kwargs: PL’s TimeSeriesDataModule args, see documentation.*

​
Usage Example

import pandas as pd
import matplotlib.pyplot as plt

from neuralforecast import NeuralForecast
from neuralforecast.losses.pytorch import DistributionLoss
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

model = TimesNet(h=12,
                 input_size=24,
                 hidden_size = 16,
                 conv_hidden_size = 32,
                 loss=DistributionLoss(distribution='Normal', level=[80, 90]),
                 scaler_type='standard',
                 learning_rate=1e-3,
                 max_steps=100,
                 val_check_steps=50,
                 early_stop_patience_steps=2)

nf = NeuralForecast(
    models=[model],
    freq='ME'
)
nf.fit(df=Y_train_df, static_df=AirPassengersStatic, val_size=12)
forecasts = nf.predict(futr_df=Y_test_df)

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])

if model.loss.is_distribution_output:
    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['TimesNet-median'], c='blue', label='median')
    plt.fill_between(x=plot_df['ds'][-12:], 
                    y1=plot_df['TimesNet-lo-90'][-12:].values, 
                    y2=plot_df['TimesNet-hi-90'][-12:].values,
                    alpha=0.4, label='level 90')
    plt.grid()
    plt.legend()
    plt.plot()
else:
    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['TimesNet'], c='blue', label='Forecast')
    plt.legend()
    plt.grid()