DeepStateOptions<T>

Options & Configuration · AiDotNet.Models.Options

Configuration options for the DeepState (Deep State Space) model.

Inheritance: ModelOptions → RegressionOptions<T> → TimeSeriesRegressionOptions<T> → DeepStateOptions<T>

For Beginners

DeepState is like giving a classical statistics model a brain:

What is a State Space Model? SSMs assume your data has hidden “states” that evolve over time:

  • State transition: z_t = F * z_{t-1} + noise (how states evolve)
  • Observation: y_t = H * z_t + noise (how states produce observations)

Example - Trend + Seasonality: States might represent:

  • Level (current baseline)
  • Trend (direction of change)
  • Seasonal patterns (weekly/yearly cycles)

The observed value is a combination of these hidden states.

Why “Deep” State Space? Classical SSMs need you to specify the model structure (how many seasonal patterns, etc.). DeepState uses a neural network to:

  • Automatically learn appropriate state representations
  • Adapt to complex, non-linear patterns
  • Share patterns across multiple time series

Benefits:

  • Interpretable decomposition (trend, seasonality, residual)
  • Natural uncertainty quantification
  • Handles multiple related time series well
  • Works with irregular data and missing values

How It Works

DeepState combines deep learning with classical state space models (SSM) for probabilistic time series forecasting. A neural network learns to parameterize the state space model, while the SSM structure provides interpretable components like trend and seasonality.

Reference: Rangapuram et al., “Deep State Space Models for Time Series Forecasting”, 2018. https://papers.nips.cc/paper/2018/hash/5cf68969fb67aa6082363a6d4e6468e2-Abstract.html

Constructors

Constructor Summary
DeepStateOptions() Initializes a new instance of the DeepStateOptions class with default values.
DeepStateOptions(DeepStateOptions<T> other) Initializes a new instance by copying from another instance.

Properties

Property Summary
bool UseSeasonality { get; set; } Gets or sets whether to include seasonality components.
bool UseTrend { get; set; } Gets or sets whether to include a trend component.
double DropoutRate { get; set; } Gets or sets the dropout rate for regularization.
double LearningRate { get; set; } Gets or sets the learning rate for training.
int BatchSize { get; set; } Gets or sets the batch size for training.
int Epochs { get; set; } Gets or sets the number of training epochs.
int ForecastHorizon { get; set; } Gets or sets the forecast horizon (prediction length).
int HiddenDimension { get; set; } Gets or sets the hidden dimension of the RNN encoder.
int LookbackWindow { get; set; } Gets or sets the lookback window size (input sequence length).
int NumRnnLayers { get; set; } Gets or sets the number of RNN layers.
int StateDimension { get; set; } Gets or sets the state dimension of the state space model.
int[] SeasonalPeriods { get; set; } Gets or sets the seasonal periods to model.