1
    

2
    know linear regression very well

3
 make sure you know R^2 and RSE and so on statistics

    MSE, is this related to something called D^2? might have misheard the prof with that one. 



4
    given some data set: what method to use?
    quantitative or qualitative? --> then what method specifically? need to explain why to choose this method. bias-variance tradeoff; will the bias be high or low, and why?  

5
    bootstraps, cross-examination
6
    uneven classes (look at for example LDA on Credit Data)

7
    PCR and PCA is very useful but probably less important for the exam