Manifold Learning and Dimensionality Reduction for Data Visualization
summary
Compares concrete manifold-learning and dimensionality-reduction algorithms (PCA, Isomap, Spectral Embedding, t-SNE) using scikit-learn implementations and runnable Jupyter notebooks. Shows which methods preserve local neighborhoods versus global structure on digit and image datasets, demonstrates a feature-engineering pipeline that improves classifier accuracy versus raw features, and provides linked slides and notebooks so methods, parameters and code are reproducible.
AI-written summary · about
chapters (11)
- 0:00Manifold learning intro
- 2:26Topological concepts
- 6:05Neighbor-preserving maps
- 10:15Local features
- 13:34Dimensionality reduction
- 17:20Cluster analysis
- 23:03Practical examples
- 28:14Classification training
- 31:06Feature generation
- 34:05Summary
- 37:15Q&A
related talks
34mDimensionally reducing data squeezing out the good stuff
37mA Bluffer's Guide to Dimension Reduction
21mA Visual Interaction Framework for Dimensionality Reduction Based Data Exploration
40mTopological Techniques for Unsupervised Learning
54mThe Sparse Manifold Transform
43mVisualising Machine Learning Models
40mA Practical Guide to Dimensionality Reduction Techniques
02h 23mEffective Data Visualization