RQOBJECTIVE(y; d; p; t) returns the objective function from fitting a quantile linear regression with a response variate y, a design matrix d, a probability value specified by the scalar p, and using a tolerance defined by the scalar t. If the fourth argument is omitted, a default tolerance of 10-12 is used. For more details of quantile regression, see the FRQUANTILES directive. The objective is calculated as
SUM( r * (p - (r.LT.0)) )
where r is the variate of residuals from the fit.