Skip to main content
20.10: What Have We Learned?
- Last updated
Save as PDF
- Networks come in various types and can be represented in probabilistic and algebraic views
- Different centrality measures gauge the importance of nodes/edges from di↵erent aspects
- PCA and SVD are useful for uncovering structural patterns in the network by performing matrix decomposition
- Sparse PCA improves upon PCA by selecting a few most representative variables in the data and more accurately recovers community structure
- Network communities have a variety of definitions, each of which has specific algorithms designed for community detection
- Neural networks and deep learning networks are supervised learning machines that capture complex patterns in data.