ABSTRACT
In the case of the classical linear regression model, the risk score is an estimate for
η(x1,x2, . . . , xp) = β0 +β1x1 +β2x2 + . . . +βpxp and this is nothing else but μ(x1,x2, . . . , xp), that is, the expected value of our outcome variable Y given the covariate values x1,x2, . . . , xp of a subject. In other words, the true value η(x1,x2, . . . , xp) is the value of Y we can expect on average if we have many subjects all with the same covariate values x1,x2, . . . , xp. And our risk score ηˆ(x1,x2, . . . , xp) is an estimate for this value.
