Deep learning networksΒΆ

aeon networks are the models behind the deep learning estimators, and can be used in

their own right to construct bespoke solutions.

BaseDeepLearningNetwork([soft_dependencies, ...])

Abstract base class for deep learning networks.

TimeCNNNetwork([n_layers, kernel_size, ...])

Establish the network structure for a CNN.

EncoderNetwork([kernel_size, n_filters, ...])

Establish the network structure for an Encoder.

FCNNetwork([n_layers, n_filters, ...])

Establish the network structure for a FCN.

InceptionNetwork([n_filters, ...])

Inception Network.

MLPNetwork([n_layers, n_units, activation, ...])

Establish the network structure for a MLP.

ResNetNetwork([n_residual_blocks, ...])

Establish the network structure for a ResNet.

AEFCNNetwork([latent_space_dim, ...])

Establish the network structure for a AE-FCN.

AEResNetNetwork([latent_space_dim, ...])

Establish the network structure for a AE-ResNet.

LITENetwork([use_litemv, n_filters, ...])

LITE and LITE Multivariate (LITEMV) Networks.

AEBiGRUNetwork([latent_space_dim, n_layers, ...])

A class to implement an Auto-Encoder based on Bidirectional GRUs.

DisjointCNNNetwork([n_layers, n_filters, ...])

Establish the network structure for a DisjointCNN Network.

DCNNNetwork([latent_space_dim, n_layers, ...])

Establish the network structure for a DCNN-Model.

AEDCNNNetwork([latent_space_dim, ...])

Establish the Auto-Encoder based structure for a DCN Network.

AEAttentionBiGRUNetwork([latent_space_dim, ...])

A class to implement an Auto-Encoder based on Attention Bidirectional GRUs.

AEDRNNNetwork([latent_space_dim, ...])

Auto-Encoder based Dilated Recurrent Neural Networks (DRNN).

RecurrentNetwork([rnn_type, n_layers, ...])

Implements a Recurrent Neural Network (RNN) for time series forecasting.

NBeatsNetwork([horizon, stacks, ...])

Implementation of the N-BEATS network architecture.

TCNNetwork([n_blocks, kernel_size, dropout])

Temporal Convolutional Network (TCN) for sequence modeling.

DeepARNetwork([lstm_units, dense_units, dropout])

DeepAR Network for probabilistic time series forecasting.