1. The rank QR (x) of the x matrix can be computed $rank, if not full-rank, indicating that Xi can be represented by a linear combination of the other x's; 2. The condition number kappa (x) can also be calculated, as said upstairs, k<100, indicating a lesser degree of collinearity; if 100<k< 1000, there are more multi-collinearity, if k>1000, there are serious multiple collinearity. You can do a stepwise regression, with the step () command, such as your first model is FM=LM (), STEP (FM) can be 3. The variance expansion factor (VIF) can be used
Library (CAR) vif (Lm.sol)
The variance expansion factor of each coefficient is obtained, when 0<=vif<100, there is a strong multiplicity of collinearity, when vif>=100, multiple collinearity is very serious. This method is more commonly used!
Multiple collinearity discrimination in R language
