| A new and rapid multicomponent analysis of foods has been developed by using near infrared reflectance spectrometry. The spectroscopic technique of Near Infrared Reflectance Analysis (NIRA) has required the use of transformation of the correlation matrix, a mathematical technique of multiple regression analysis in order to generate the empirical calibration required for the success of the technique. Preliminary calibration curves for beef, pork, and beef and pork mixture have been developed using an InfraAlyzer 400('TM) (Technicon Industrial Systems, Tarrytown, N.Y.) and a Hewlett-Packard 9815'A' calculator (Hewlett-Packard Company, Santa Clara, CA). The accuracy of these preliminary prediction regression equations was not satisfactory for the use of in-process control in meat processing. In subsequent experiments, a Vita Mix blender (Vita Mix Corp., Cleveland, Ohio) was introduced to improve the homogeneity of the samples and a Hewlett-Packard 85 personal computer (Hewlett-Packard Company, Santa Clara, CA) replaced the less powerful Hewlett-Packard 9815'A'. Nineteen beef samples with varying amounts of collagen were prepared and measured via the InfraAlyzer 400('TM) and hydroxyproline. The multiple correlation coefficients (R('2)) of the prediction equations for hydroxyproline, protein, fat, and moisture were 0.9389, 0.7745, 0.9329, and 0.9425, respectively. Correlations of 0.2754, 0.5279, 0.8020, and 0.8745 were found between near infrared values and chemical analysis values for hydroxyproline, protein, fat, and moisture. The R('2) values for emulsified pork were 0.9469 for protein, 0.9662 for fat, and 0.9524 for moisture. Correlations between near infrared values and chemical analysis values were 0.8857 for protein, 0.9725 for fat, and 0.9628 for moisture. With the improved accuracy of prediction equations through the improved sampling procedure and the most capable computer programming, the near infrared reflectance analysis can be an alternative technique for rapid determination of selected meat components. |
