Discriminant and principal component analyses to classify commercial orange juices based on relative amounts of volatile juice constituents

Pasteurised orange juices and two types of commercial orange juice from fro zen concentrate were differentiated based on the combined levels of 15-25 v olatile constituents monitored in each of 44 juice samples. Multivariate an alysis programs were used to calculate discriminant and principal component analyses and to display the results graphically. Since this group of sampl es was part of a nutritional database study, other parameters were included in an effort to improve discrimination of the types of juice samples monit ored. Addition of degrees Brix and acid values improved the separation by p rincipal component analysis of the three types of juice studied, Discrimina nt analysis provided better separation of juice types than did principal co mponent analysis.