How do concepts like R-squared, multicollinearity, and statistical tests (T-test vs. F-test) help evaluate linear regression models and interpret their results effectively?