Publisher's Synopsis
The statistical dependence techniques are characterized by the fact of that one (or more) variables in study stands out as a dependent main. This concept is as opposed to statistical interdependence techniques in which no variable stands out as a dependent. In the case of the methods of the unit it is necessary to use multivariate analytical techniques or inferential whereas the dependent variable as explained by the other independent variables explanatory, and trying to relate all the variables by means of a possible equation or model that the link (for example, the regression model that generalizes the canonical correlation to several dependent variables). Once configured the mathematical model can be to predict the dependent variable (or variables) value called the profile of other. If the dependent variable qualitative dichotomous can be used as sorting machine, studying its relation with the other variables Tableau (logistic regression). The observed qualitative dependent variable found the allocation of each individual to previously defined groups (two or more than two), can be used to classify new cases in which unknown the group that probably belong (discriminant analysis), which solves the problem of allocation on the basis of a quantitative profile of Tableau variables. If the dependent variable is quantitative and the explanatory are qualitative we have models of the analysis of variance, which can extend to models loglinear analysis tables contingency of high dimension: As in all technical dependence underlies a model, usually associated these techniques to econometric models. Econometric models (methods of dependence) underlies a general relationship between the independent variables and the dependent of the generic type . The nature of the variables will characterize each model. This book develops all these models.
