Classification of Cow’s, Buffalo’s and Goat’s Milk by Fourier Transform Near Infrared Spectroscopy

บทคัดย่อ

This research aims to apply the Fourier transform near infrared spectroscopy (FT-NIRS) for classificationcow, goat and buffalo milk. Eleven milk samples included raw buffalo milk, pasteurized buffalo milk, raw cow milk, boiled cow milk, pasteurized cow milk, no fat pasteurized cow milk, sterilize cow milk, UHT cow milk, raw goat milk, pasteurized goat milk and UHT goat milk. Classification models were developed by Principal component analysis (PCA), Soft independent modeling by class analogy (SIMCA) and Partial least squares
discriminant analysis (PLS-DA) method. Classification model of raw buffalo milk, which created by SIMCA method, cloud well identify other types of milk. Classification results on models of no fat pasteurized cow milk, boiled cow milk and pasteurized cow milk, which were developed by PLS-DA based on cross-validation method, showed high percent of correct identification. Moreover, good classification models based on test set validation method were models of no fat pasteurized cow milk, boiled cow milk, sterilize cow milk, raw goat milk and pasteurized goat milk. The findings point to possible application of FT-NIRS for classification the types of milk.