We present here a novel technique for the identification of the parameters
of the probability function used in the Preisach-based modeling. This techn
ique seems to overpass the problems linked with the use of a product of two
single-valued functions as a probability function that in the case of soft
magnetic materials may cause an unacceptable lack in accuracy of the model
. Experimental tests are presented in the final section of the paper indica
ting the satisfactory improvement of the predictive capacity of the model d
ue to the proposed identification technique.