Explore MATLAB source code curated for "非线性降维" with clean implementations, documentation, and examples.
Locality Preserving Projections (LPP) as a nonlinear dimensionality reduction method can be effectively applied to machine learning with high-dimensional datasets using graph-based neighborhood preservation techniques.
This program implements Locally Linear Embedding (LLE), one of the representative nonlinear dimensionality reduction algorithms, which is valuable for researchers focusing on high-dimensional data reduction studies. The implementation demonstrates the algorithm's core workflow through practical code examples and data processing steps.
A manifold learning algorithm for nonlinear dimensionality reduction, originally published in Science magazine in 2000, with enhanced code implementation details and algorithm explanation
Laplacian Eigenmaps is a manifold learning-based nonlinear dimensionality reduction technique that constructs weights using heat kernels, widely applicable in image segmentation to enhance clustering performance through spectral graph theory implementation.
A manifold learning algorithm for nonlinear dimensionality reduction that primarily represents geometric features at each point using local tangent spaces and then aligns these local tangent spaces for global embedding.
Introduction to Locality Preserving Projections (LPP) as a nonlinear dimensionality reduction technique, with algorithm principles and implementation insights
A detailed implementation code for Kernel PCA with algorithmic explanations and key function descriptions