There are several Multivariate Data Analysis techniques that are used in the development of research, with well-defned objectives, in the area of tourism. In fact, most of these techniques fall within the scope of data classifcation, dimensionality reduction and, above all, the detection of interesting patterns in databases through graphical factorial methods. Typically, factorial methods present the results in the form of scatter diagrams, generally in a twodimensional subspace, even if the original confguration is larger. The frst problem to be taken into account is the number of dimensions required to obtain an adequate representation in reduced dimension. However, since the sequential obtaining of each of the axes of the representation is identical to that obtained from the joint adjustment of all of them, it is possible to choose the number of axes necessary after performing the singular value decomposition calculation. In the literature there are several procedures that allow the search for the number of dimensions required to optimally describe the points im the cloud. The methods are initially described for Principal Components or Correspondence Analyses, but can be extended, according to Galindo & Cuadras , to BIPLOT methods. 21 Indeed, among Multivariate Data Analysis methods, BIPLOT methods have been the source of continuous contributions to science since their origin with Gabriel in 1971 being, however, still little known in the tourism feld. A BIPLOT is a graphical representation of multivariate data. In the same way that a scatter diagram describes the joint distribution of two variables, a BIPLOT represents three or more variables. The BIPLOT approximates the distribution of a multivariate sample to a reduced dimensional space, usually of dimension two, and superimposes on the same representation the variables on which the sample is measured. The representations of the variables are usually vectors and coincide with the directions that best describe the individual changes of each variable. The prefx "BI" refers to the overlap, in the same representation, of individuals and variables. BIPLOTs are useful to graphically describe data or to show the results of more formal models and, from the user's point of view, they are important because their interpretation is based on simple geometric concepts that are part of mathematical culture, namely: the similarity between individuals is an inverse function of the distance between them; the lengths and angles of the vectors representing the variables are interpreted in terms of variability and covariability, respectively; the relationships between individuals and variables are interpreted in terms of the scalar product, i.e. in terms of the projections of the "individual" points onto the "variable" vectors. Galindo, in 1986, proposes what she calls HJ-BIPLOT. In this representation, the interpretation of individuals and variables is the same, without, however, preventing the search for the variables that determine the diferences between individuals from being carried out through factorial axes, that is, the new variables are interpreted as linear combinations of the initial variables and their relationships with the observed variables are interpreted. The measure of the relationship between the axes of the HJ-BIPLOT representation and each of the observed variables is called the Relative Factor Contribution to the Element (variable), which represents the part of the variability of each of the variables explained by the factor and is interpreted in the same way as a coefcient of Galindo & Cuadras, 1986. 21 17
RkJQdWJsaXNoZXIy MTE4NzM5Nw==