Análisis discriminante canónico con técnicas gráficas multivariadas aplicado a un diseño con dos factores
Autor: | Javier Lorbes Medina, Lisbeth Díaz, Manuel Milla Pino, Yelitza García Orellana |
---|---|
Rok vydání: | 2014 |
Předmět: | |
Zdroj: | Avances Investigación en Ingeniería. 11:38-47 |
ISSN: | 2619-6581 1794-4953 |
Popis: | In this paper two are described graphics techniques for multivariate data in the context of canonical discriminant analysis applied to a multivariate design with two factors and graphic representations of results generated application. In particular, the methodology described and illustrated to provide a low dimensional view data obtained from a design with two factors based multivariate canonical discriminant analysis with graphic data in reduced for multiple responses such as graphical range Hypothesis-Error (HE), which provides a direct visual comparison of the covariance matrices for the hypothesis and error, and the graph of canonical discriminant structure, which shows an alternative view for all variables in two-dimensional space that maximizes the differences between the groups, which provide a compact visual summary of the salient features of the results as they show all observations, group means, and their relationships to varying responses. In one application, the scope and potential of the canonical discriminant analysis with these graphs as an alternative for complex data analysis demonstrated experimental designs. |
Databáze: | OpenAIRE |
Externí odkaz: |