ANOVA like analysis for structured families of stochastic matrices
Cristina Dias, Carla Santos, Maria José Varadinov, João Tiago Mexia · AIP conference proceedings · 2016
Symmetric stochastic matrices width a width a dominant eigenvalue λ and the corresponding eigenvector α appears in many applications. Such matrices can be written as M=λ α αt+E¯.Thus β = λ α will be the structure vector. When the matrices in such families correspond to the treatments of a base design we can carry out a ANOVA like analysis of the action of the treatments in the model on the structured vectors. This analysis can be transversal–when we worked width homologous components and - longitudinal when we consider contrast on the components of each structure vector.The analysis will be briefly considered at the end of our presentation.