Parametric inference for proportional (reverse) hazard rate models with nomination sampling