Abstract: In order to decrease the rejection ratio in the rubber-producing process‚the dispersed homogeneous degree of carbon-black in rubber should be indirectly‚real time and accurately determined．Support Vector Machines based on the principle of structural risk minimization is a new machine learning method which is of excellent classification and generalization ability when being used in the small sample decision．In this paper‚a6-band determination model of dispersed homogeneous degree of carbon-black in rubber was established on the basis of mult-i class Least Squares Support Vector Machines‚and some practical data were determined by the model．The results show that the proposed model is practical and average rate of false determination reduce to3．6％‚and can determine the dispersed homogeneous degree of carbon-black in rubber quickly and correctly．

Key words: carbon-black, dispersed homogeneous degree, least squares support Vector Machine, discrimination model