SparseGaussianProcess<T>
Models & Types · AiDotNet.GaussianProcesses
A sparse implementation of Gaussian Process regression that uses inducing points to reduce computational complexity.
Inheritance: ModelBase<T, Matrix<T>, Vector<T>> → GaussianProcessBase<T> → SparseGaussianProcess<T>
Implements: IGeneratedParameterRegistrar<T>
For Beginners
A Gaussian Process is a flexible machine learning method that can model complex relationships in data.
The “sparse” version solves a common problem with Gaussian Processes - they can be very slow with large datasets. Instead of using all data points for predictions (which can be computationally expensive), this implementation selects a smaller set of representative points called “inducing points” that capture the essential patterns in your data.
Think of it like summarizing a book: instead of reading every word (standard GP), you read just the chapter summaries (sparse GP) to get the main ideas more efficiently.
This approach makes Gaussian Processes practical for larger datasets while maintaining most of their predictive power.
Constructors
| Constructor | Summary |
|---|---|
SparseGaussianProcess(IKernelFunction<T> kernel, MatrixDecompositionType decompositionType, double noiseVariance) |
Methods
| Method | Summary |
|---|---|
ValueTuple<T, T> Predict(Vector<T> x) |
Makes a prediction for a new input point, returning both the mean prediction and its variance. |
void Fit(Matrix<T> X, Vector<T> y) |
Trains the Gaussian Process model on the provided data. |
void UpdateKernel(IKernelFunction<T> kernel) |
Updates the kernel function used by the model and retrains if data is available. |