# DFBETA

## Detecting Influential Points in Regression with DFBETA(S)

In regression modeling, influential points are observations that, individually, exert large effects on a model’s results—the parameter estimates ($$\hat{\beta_0}, \hat{\beta_1}, …, \hat{\beta_j}$$) and, consequently, the model’s predictions ($$\hat{y_1}, \hat{y_2}, …, \hat{y_i}$$). Influential points aren’t necessarily troublesome, but observations flagged as highly influential warrant follow-up. A large value on an influence measure can signal anything from […]