SRM β Statistics for Risk Modeling
SOA
Coefficient of determination , adjusted
Logistic regression ,
Biasβvariance
PCA proportion of variance of component :
Exponential smoothing
Traps to remember
Adjusted can fall when adding a predictor; never does.
Interpreting a logistic coefficient as a change in probability instead of log-odds.
Confusing k-fold CV with LOOCV bias/variance properties.
Assuming random forests need pruning (they do not).