In many real physical systems based fuzzy inference systems the rulebase is sparse thus interpolation or the change of the shape of the rules become necessary if the rulebase parameters are selected according to physical parameters of the systems. Often measurements contain noise and outlines which can draw the statistics of the measured data. In the present article based on two independent examples, namely telecommunication line evaluation and colonoscopy image processing, we study the effect of the selection of the rulebase parameters on the effectiveness of stabilized fuzzy KH interpolation.