A data matrix in which pixel numbers are assigned to the vertical direction, m/z values are assigned to the horizontal direction, and intensity values are used as terms is generated from data obtained by peak detection (S3), and after the standardization of the data is executed so that the norms of the intensities of the pixel space are set to 1 for each m/z, the peaks (m/z values) are classified into a plurality of clusters by performing clustering in the m/z direction (S5 and S6). Since the probability of isotope peaks or adduct ion peaks derived from the same substance being consolidated into the same cluster increases, unnecessary peaks can be accurately removed by removing unnecessary isotope ion peaks and the like using m/z differences or intensities in the clusters (S7 and S8).