Spectrum data of white neutrons having different spectrum shapes, and SEE counts obtained by a white neutron method using this multiple spectrum data, are stored. A computing section reads out the spectrum data, divides the data into energy groups, and calculates and stores a total flux of each energy group. Furthermore, the computing section reads out the SEE counts with respect to each of the spectrum data and the total flux of each energy group, substitutes the SEE counts and the total flux into a simultaneous equation, and calculates the SEE cross section for each of the energy groups. Subsequently, the computing section calculates parameters which determine a formula of the approximate function of the SEE cross section as a function of energy, so that computed values of error counts obtained by integration of multiple spectra and the approximate function sufficiently match the actual measured values thereof.