Multivariate analysis

Abstract

This chapter discusses multivariate analysis. The term multivariate could be applied to a number of the problems. Multivariate analyses may be exploratory, seeking out pattern or structure in a single set of variables, or they may be directed to the estimation and use of relationships between two or more such sets of variables. The precise nature of the investigation of multivariate methods in general is quite complex, both in terms of their theoretical background and in terms of computational demands. The linear discriminate function is a technique for classifying an object or individual into one of two possible groups on the basis of a set of observations or measurements. An unbiased estimate of the actual error rate that is implicit in the use of a given classification rule can be obtained by applying the rule to an independent sample of subjects of known classification status. This procedure is referred to as cross-validation.