The dependence of the diameter swell and sag of a parison extruded from blends of high-density polyethylene (HDPE)/polyamide-6 (PA6)/compatibilizer on the blend composition and flow rate is determined by analyzing video images of the parison. Among the blends tested, blends with a PA6 concentration of 35% or below exhibit an appreciable swell. A greater degree of sag begins to appear for the blend with a PA6 content of 45%. A neural network approach is applied to the experimental data, leading to a model for predicting the diameter swell profile from new levels of input variables, including the blend composition and flow rate.