Journal of Food Science · 1999 · 63 citations · 13 references
AgricultureAnimal SciencesFood ScienceMeat ScienceSpectroscopic AnalysisSpectroscopyInfrared SpectroscopyMeat ClassificationMeat CompositionMeat QualityNear-infrared SpectroscopySpectroscopic Food AnalysisBeef ClassificationClassification AccuracyClassification Accuracy AssessmentLinear Discriminant AnalysisMeat Species IdentificationMultiple RegressionReflectanceStepwise Regression
ABSTRACT: A visible/near‐infrared (NIR) spectroscopic method was developed with canonical discriminant analysis and stepwise multiple linear regression to differentiate beef from kangaroo meat. Results showed that beef and kangaroo meat could be separated when samples were analyzed by spectrophotometry as minced meat or cut meat. For minced meat, scatter correction and derivative treatment of reflectance spectra improved classification. For cut meat, original reflectance spectra produced better classification. Overall classification accuracy was 83% to 100%, and no kangaroo meat was misclassified. NIR spectroscopy might be an efficient test method for species identification.
13