When a user inputs samples per group, a sample tree and a peak matrix are generated. Peak lists per group are shown in the sample tree, and m/z values and signal strength values from the peak lists are coordinates in the peak matrix. A multivariate analysis is applied to the generated peak matrix. The sample tree, peak matrix, score plot, and loading plot are displayed. When the user clicks a plotted point on the loading plot, a row indicating a corresponding peak on the peak matrix is discriminated. When the user deletes a checkmark corresponding to the discriminated row, the multivariate analysis is applied to the peak matrix from which the peak has been excluded. The score plot and other data are updated. When separation between groups is known from the score plot as failure, the excluded peak may be visually determined as a marker contributing to the group separation.