TFCOptions<T>
Options & Configuration · AiDotNet.Models.Options
Configuration options for TF-C (Time-Frequency Consistency for Self-Supervised Time Series).
Inheritance: ModelOptions → RegressionOptions<T> → TimeSeriesRegressionOptions<T> → TFCOptions<T>
How It Works
TF-C learns time series representations by enforcing consistency between time-domain and frequency-domain representations via contrastive learning, capturing both temporal and spectral patterns.
Reference: Zhang et al., “Self-Supervised Contrastive Pre-Training For Time Series via Time-Frequency Consistency”, NeurIPS 2022.
Constructors
| Constructor | Summary |
|---|---|
TFCOptions() |
Creates options with the default values. |
TFCOptions(TFCOptions<T> other) |
Creates a copy of another options instance. |
Properties
| Property | Summary |
|---|---|
double ContrastiveTemperature { get; set; } |
Temperature of the contrastive (NT-Xent) loss. |
double DropoutRate { get; set; } |
Dropout probability applied during training. |
int ContextLength { get; set; } |
Number of past time steps the model reads as context. |
int ForecastHorizon { get; set; } |
Number of future time steps to forecast. |
int HiddenDimension { get; set; } |
Width of the encoders’ hidden representations. |
int NumFreqLayers { get; set; } |
Number of layers of the frequency-domain encoder. |
int NumTimeLayers { get; set; } |
Number of layers of the time-domain encoder. |
int ProjectionDimension { get; set; } |
Size of the shared time-frequency space both encoders project into. |