Walmart Labs Interview Question | Predictors
Question
If two predictors are highly correlated, what is the effect on the coefficients in the logistic regression?
in progress
1
Statistics
4 years
3 Answers
902 views
Great Grand Master 0
Answers ( 3 )
Because the variance of the sampling distribution of the regression coefficient would be larger (by a factor of the VIF) if it were correlated with other variables in the model, the p-values would be higher (i.e., less significant) than they otherwise would.
The variance of the coefficient estimate increases and the model can give varied results
with the small change in the data. This makes the model very unstable.
We can no longer make much sense of the usual interpretation of a slope coefficient as the
change in the mean response for each additional unit increase in the predictor xk, when all the
other predictors are held constant.
The equation for logit is (1/1+e^-Z)value where Z=a+B1*x1+B2*x2+B3*x3…Bn*xn
when variables are highly correlated the value of coefficient tends to increase and it leads to instability.
When we have this in 1/(1+e-Z ) it tends to minimize the value more than usual and thus leads to insatiability in classification