NHiTSModel<T>

Models & Types · AiDotNet.TimeSeries

Implements N-HiTS (Neural Hierarchical Interpolation for Time Series) for efficient long-horizon forecasting.

Inheritance: TimeSeriesModelBase<T> → NHiTSModel<T>

Implements: IGeneratedParameterRegistrar<T>, ISupportsLossFunction<T>

For Beginners

N-HiTS improves upon N-BEATS by using a “zoom lens” approach to time series. It looks at your data at three different zoom levels:

  • Zoomed out (low resolution): Captures long-term trends like yearly seasonality
  • Medium zoom: Captures medium-term patterns like monthly cycles
  • Zoomed in (high resolution): Captures short-term fluctuations like daily variations

By combining insights from all three levels, it produces more accurate forecasts, especially for predicting far into the future.

How It Works

N-HiTS is an evolution of N-BEATS that addresses limitations in long-horizon forecasting through:

Original paper: Challu et al., “N-HiTS: Neural Hierarchical Interpolation for Time Series Forecasting” (AAAI 2023).

Production-Ready Features:

  • Uses Tensor for GPU-accelerated operations via IEngine
  • Proper backpropagation via automatic differentiation
  • Vectorized operations - no scalar loops in hot paths
  • All parameters are trained (not subsets)

Example

using AiDotNet;
using AiDotNet.Data.Loaders;
using AiDotNet.TimeSeries;
using AiDotNet.Tensors.LinearAlgebra;

double[] series =
{
    120, 135, 148, 160, 155, 170, 180, 195, 210, 198, 220, 235,
    140, 155, 165, 178, 172, 190, 200, 215, 230, 218, 245, 260
};
var x = new Matrix<double>(series.Length, 1);
for (int i = 0; i < series.Length; i++) x[i, 0] = i;

var result = await new AiModelBuilder<double, Matrix<double>, Vector<double>>()
    .ConfigureModel(new NHiTSModel<double>())
    .ConfigureDataLoader(DataLoaders.FromMatrixVector(x, new Vector<double>(series)))
    .BuildAsync();

var forecast = result.Predict(new Matrix<double>(6, 1));
Console.WriteLine($"NHiTSModel: forecast {forecast.Length} steps.");

Constructors

Constructor Summary
NHiTSModel(NHiTSOptions<T> options) Initializes a new instance of the NHiTSModel class.

Methods

Method Summary
ModelMetadata<T> GetModelMetadata()
T PredictSingle(Vector<T> input)
Vector<T> ForecastHorizon(Vector<T> input) Generates forecasts for the full horizon using hierarchical processing.
Vector<T> Predict(Matrix<T> input)
void SetLossFunction(ILossFunction<T> lossFunction)