Classification Modelling of Data Structures

The basis of classification is supervised learning where a set of known objects that belong unambiguously to certain classes are analyzed. From their features (analytical data) classification rules are obtained by means of relevant properties of the data like dispersion and correlation. 

One of the powerful classification methods is multivariate variance and discriminant analysis (MVDA) (Dillon and Goldstein [1984] Ahrens and Lauter [1974] Danzer et al. [1984]). 

By means of eigenanalysis multivariate discriminant functions, df, can be derived with eigenvectors Vj (j = 1,2.p) where p m is the rank of the matrix R see Eq. (8.14), m is the number of original variables (i = 1,2. m). With Eq. (8.16) the discriminant functions are linear combinations of the original variables 

80-95% of the total information and, therefore, a two-dimensional plot of them may represent a realistic illustration of the respective classification as is schematically shown in Fig. 8.10. The boundaries between the classes may be estimated according to 

In a corresponding way, measures of other class figures, e.g., ellipses, can also be calculated. 

---