At work, I predict insurance (auto and homeowners) losses based on customer characteristics. We have a dependent variable (losses), and various predictors (e.g., age, gender, marital status, etc.).
Age and marital status are correlated: older people are more likely to be married. If you take an extremely simplistic approach to forming predictions
(
Read more... )