Data can be multidimensional. However, we usually reshape those data in matrices and vectors and then analyze them. This approach, simplifying the tractability of the object, destroys its interconnections in the multidimensional space.
For this motivation, I am working in the field of High dimensional data that can be structured in multidimensional matrices. These data are referred to as Multiway data or Tensors. I am currently working on the application of such structures to financial and economic data.
In particular, I proposed a tensor autoregression for different types of data such as covariance matrices and macroeconomics variables, in which the dimensions of the tensor are the variables of interest (first mode), the countries (second mode), and the time (third mode), generating a three-dimensional tensor.
