Spline-Based Mixture Distribution Analysis (C# / WinForms)

Bachelor's/specialist diploma project implementing a desktop application for statistical mixture-distribution analysis using spline surfaces, Bayesian classification, and least-squares estimation.

What this is

A WinForms desktop app (diplom2_0/) that models 2D data as a mixture of distributions and classifies/segments it using:

• Spline.cs — spline surface fitting over the data density
• Bayes.cs — Bayesian classification between mixture components
• MNK.cs — least-squares (МНК) parameter estimation
• MGK.cs — mixture/Gaussian-component computations
• LocalMax.cs — local maxima (mode) detection on the fitted surface
• MixDistribution.cs / Linears.cs / Line.cs — mixture and linear-boundary modeling
• Draw.cs / Window.cs — visualization of the fitted surface and classification result

Why it's here

This project is the direct predecessor of my PhD dissertation research: the same core ideas (spline-based density modeling, mode detection, discriminant classification) were later reformulated, generalized, and reimplemented in Python/NumPy and PyTorch:

• spline-analyzer-dissertation — NumPy research framework
• spline-analyzer-torch — PyTorch reimplementation

Stack

C# / .NET Framework, WinForms.