A simple form of multivariate data is a rectangular table (matrix, spreadsheet) consisting of n rows, m columns, and each cell containing a numerical value. Each row corresponds to an OBJECT, for instance a sample; each column corresponds to a particular FEATURE of the objects (VARIABLE, for instance a measurement on the objects). We call these data the matrix X, with element xij in row i and column j. A column vector xj contains the values of variable j for all objects; a row vector, xTi , is a transposed vector and contains all features for object i (Table 2.1).