BSplineKernel<T>

Models & Types · AiDotNet.Kernels

Implements the B-Spline kernel function for measuring similarity between data points.

Implements: IKernelFunction<T>

For Beginners

A kernel function is a mathematical tool that measures how similar two data points are. The B-Spline kernel is a specialized similarity measure that works well for data that needs smooth transitions between points. Think of B-splines like drawing a smooth curve through a set of points, where the curve doesn’t necessarily pass through all points but creates a smooth approximation.

How It Works

The B-Spline kernel is based on B-spline functions, which are piecewise polynomial functions with compact support. This kernel is particularly useful for problems requiring smooth interpolation and approximation.

This kernel is particularly useful when you want your AI model to produce smooth outputs and avoid abrupt changes in predictions, similar to how a skilled artist might draw a smooth curve rather than connecting dots with straight lines.

Example

using AiDotNet.Kernels;
using AiDotNet.Tensors.LinearAlgebra;

var kernel = new BSplineKernel<double>();
var a = new Vector<double>(new[] { 1.0, 2.0, 3.0 });
var b = new Vector<double>(new[] { 1.5, 1.0, 2.5 });
double similarity = kernel.Calculate(a, b);
Console.WriteLine($"BSplineKernel: K(a, b) = {similarity:F4}");

Constructors

Constructor Summary
BSplineKernel(int degree, T knotSpacing) Initializes a new instance of the B-Spline kernel with optional parameters.

Methods

Method Summary
T Calculate(Vector<T> x1, Vector<T> x2) Calculates the B-Spline kernel value between two vectors.