Bayesian scalar-on-tensor regression using the Tucker decomposition for sparse spatial modeling.

Modeling with multidimensional arrays, or tensors, often presents a problem due to high dimensionality. In addition, these structures typically exhibit inherent sparsity, requiring the use of regularization methods to properly characterize an associa…

Preview only. Read the full abstract at the source