To calculate the degree of coincidence of an unknown spectrum with a known spectrum accurately by obtaining a correlation between the known and unknown spectra at the peak position according to the differential spectrum of the known spectrum and calculating the coincidence of the unkown spectrum with the known spectrum on the basis of its correlation coefficient.

A data around a base line BL is delected, at a constant threshold, from a differential spectrum that is obtained through secondary differntiation of a known spectrum. The data larger than the threshold SL is left as it is, therbey determining the data points of the known and unknown spectra. Then, the known and unknown spectra are regulated so as to be included between zero and one. Further, the data corresponding to the extracted data points are allowed to exit from the regulated spectra. Then, the data of the known spectrum are shown by the horizontal axis, and those of the unknown spectrum are shown by the vertical axis, further the corresponding data are shown on a coordinate system, thereby obtaining points P1 to P6 therefrom. Thus, the correlation coefficient of two spectra data is obtained on the basis of the distribution diagram of the two spectra data.