CLASSIFICATION AND SEGMENTATION TECNIQUES. Examples with SPSS

The initial classification of segmentation techniques distinguish between predictive techniques, in which the variables that intervene in the process qualify initially in dependent and independent (similar to the analysis of dependence or explanatory methods of multivariate analysis techniques) and descriptive techniques, in which all the variables initially have the same status (similar to the analysis of interdependence or descriptive methods of multivariate analysis techniques). Predictive segmentation techniques specify the model to the data based on theoretical knowledge. The model for data should contrast after the process of data mining before accepting it as valid. Formally, the application of all model must overcome objective identification phases (from the data apply rules that permit to identify the best possible model that fits the data), estimate (process of calculation of parameters of the model chosen for the data in the identification phase), diagnosis (contrast of the validity of the estimated model process) and forecast (process model identified, estimated, and validated to predict future values of the dependent variables). We can include the predictive models all types of regression, analysis of variance and covariance, analysis of time series, mixed models, models of panel data, logit and probit models and non-linear models. All of these models will be explored throughout this book, both theoretically and SPSS applications. Exercises resolved for all chapters with the SPSS software, one of the most suitable for the job market are presented.