How does multicollinearity affect the estimation of regression parameters?

Multicollinearity will make the model estimation distorted or difficult to estimate accurately because of the exact correlation or high correlation between explanatory variables in linear regression model. The specific impacts are as follows:

1, the economic significance of the parameter estimator is unreasonable;

2. The significance test of variables is meaningless, which may exclude important explanatory variables from the model;

3. The prediction function of the model is invalid. Large variance tends to enlarge the "interval" of interval prediction, making the prediction meaningless.

Extended data

Multiple collinearity increases the variance of parameter estimation, and the greater the variance expansion factor, the stronger the collinearity. On the contrary, because admissibility is the reciprocal of variance expansion factor, the smaller the admissibility, the stronger the collinearity.

It can be remembered that permissibility represents permissibility, that is, permission. If the value is smaller, the value is not allowed, that is, the smaller it is, the less it is allowed. Collinearity is a negative indicator and is not expected to appear in the analysis. Collinearity and acceptability are linked. The variance expansion factor is the reciprocal of the tolerance, so the other way around.

Baidu encyclopedia-multicollinearity