manifold learning

Uniform manifold approximation and projection (UMAP) is a dimension reduction technique used in data science for visualizing and interpreting high-dimensional data. It is known for preserving both local and global data structure, making it valuable for exploratory data analysis and machine learning.