PCA
PCA
Dimensionality Reduction
Dimensionality Reduction Techniques
less than a minute
This course includes
3 articles
10 mins reading time
3 videos • 1 hr 31 mins
Course content
3 sections • 36 topics • Read: 10 mins • Video: 1 hr 31 mins
PCA
Use Case
Intuition
Principal Component Analysis
Goal
PCA as Optimization Problem
Why Variance of Projection is \(u^{T}\Sigma u\)?
Constrained Optimization
Constrained Optimization Solution
PCA Algorithm
Drawbacks
Principal Component Analysis (PCA) | Dimensionality Reduction | Explained with Example
t-SNE
Use Case
Intuition
t-SNE
Problem
Solution
High Dimensional Space (Gaussian)
Low Dimensional Space (t-distribution)
Optimization ️
Gradient Descent
Meaning of Terms
Interpretation
Gradient Descent Update Step
Perplexity
t-SNE Plot of MNIST Digits
t-Distributed Stochastic Neighbor Embedding ( t-SNE) | Dimensionality Reduction | Explained
UMAP
Use Case
Intuition
UMAP
Problem
Solution
High Dimensional Graph (Manifold Approximation)
Low Dimensional Space (Optimization)
Optimization
Cross Entropy Loss
Stochastic Gradient Descent
UMAP Plot of MNIST Digits
Drawbacks
Uniform Manifold Approximation and Projection (UMAP) | Dimensionality Reduction | Explained