Anova: Component Effect

The Component Effect (Ei) is the difference between a coefficient and the average of all the other coefficients. It is effectively the amount the response changes as the component goes from its low to its high setting. It is used to compute the component gradient.

\(E_i=\beta_i-\frac{\sum_{j\neq i}^{q}\beta_j}{(q-1)}\)


βi is the ith coefficient

q is the number of components