Fourier transform infrared (FTIR) and Fourier transform Raman (FT-Raman) methods were used for rapid characterization and classification of selected irradiated starch samples. Biochemical changes due to irradiation were detected using the two vibrational spectroscopic techniques, and canonical variate analysis (CVA) was applied to the spectral data for discriminating starch samples based on the extent of irradiation. The O-H (3000-3600 cm(-1)) stretch, C-H (2800-3000 cm(-1)) stretch, the skeletal mode vibration of the glycosidic linkage (900-950 cm(-1)) in both Raman and infrared spectra, and the infrared band of water adsorbed in the amorphous parts of starches (1550-1750 cm(-1)) were employed in classification analysis of irradiated starches. Spectral data related to water adsorbed in the noncrystalline regions of starches provided a better classification of irradiated starches with 5 partial least-squares (PLS) factors in the multivariate model.