So Hurrah! We've finished. Now we need to review everything.

So far we have covered:

1. Distance metrics
   • cosine similarity
   • euclidean distance (L2 norm)
   • Manhattan distance (L1 norm)
   • Chebyshev distance
   • Jaccard distance
2. collinearity and multicollinearity
3. support vector machines (SVM)
4. principle component analysis (PCA)

Still to cover:

• item-based vs. user-based recommenders. which similarity matrix do you use and why?
• cost function, gradient descent
• lasso vs. ridge
• NMF
• LDA
• MCMC
• heteroscedasticity
• a/b testing
• scoring methods
• clustering
• anamoly detection
• SNA
• Command line utilities for data cleaning
• Useful utilities for data visualization
• Tips on approaching coding questions
• graphs and network analysis, social network