Multicollinearity monitoring
- Variance inflation factor: $$ mathit{VIF_j} = frac{1}{1-mathit{R^2_j}} $$
- Correlation: $$ mathit{sCorr(x,y)} = frac{sum(x_i - hat x)(y_i - hat y)/(n-1)}{sqrt{mathit{sVar(x)ast sVar(y)}}} $$
- Indicator values: $ mathit{VIF_j}>10 $, $ mathit{sCorr(x,y)}>0.9 $