Equation 3.11 defines the matrix P , an n × n matrix determined entirely by the Xs. This matrix plays a particularly important role in regression analysis. It is a symmetric matrix (,P-'.= P ) that is also idempotent (P P = P ), and is therefore a projection matrix (see Section 2.6). Equation 3.11 shows that ,Y. is a linear function of Y with the coefficients given by . (For example, the first row of P contains the coefficients for the linear function of all ,Y-i. that gives ,Y. 1.)