Aiming at the problems of small samples, uncertainty data and heteroscedastic noise in design time forecast,
fuzzy regression theory is combined with par-Fv-SVM, the constraints are formulated based on the necessity model, and
a kind of fuzzy support vector machine named par-Fv-SVM is proposed. Then an intelligent forecast method of design
activity time and its relevant parameter-choosing algorithm are proposed. The results of application in injection mold designs
show that the time forecast method based on par-Fv-SVM is of feasibility and effectiveness.