Log In Sign Up

Soft Tensor Regression

by   Georgia Papadogeorgou, et al.

Statistical methods relating tensor predictors to scalar outcomes in a regression model generally vectorize the tensor predictor and estimate the coefficients of its entries employing some form of regularization, use summaries of the tensor covariate, or use a low dimensional approximation of the coefficient tensor. However, low rank approximations of the coefficient tensor can suffer if the true rank is not small. We propose a tensor regression framework which assumes a soft version of the parallel factors (PARAFAC) approximation. In contrast to classic PARAFAC, where each entry of the coefficient tensor is the sum of products of row-specific contributions across the tensor modes, the soft tensor regression (Softer) framework allows the row-specific contributions to vary around an overall mean. We follow a Bayesian approach to inference, and show that softening the PARAFAC increases model flexibility, leads to more precise predictions, improved estimation of coefficient tensors, and more accurate identification of important predictor entries, even for a low approximation rank. From a theoretical perspective, we show that the posterior distribution of the coefficient tensor based on Softer is weakly consistent irrespective of the true tensor or approximation rank. In the context of our motivating application, we adapt Softer to symmetric and semi-symmetric tensor predictors and analyze the relationship between brain network characteristics and human traits.


page 5

page 8

page 13

page 17

page 28

page 29

page 33


Decomposable Sparse Tensor on Tensor Regression

Most regularized tensor regression research focuses on tensors predictor...

Beyond Low Rank: A Data-Adaptive Tensor Completion Method

Low rank tensor representation underpins much of recent progress in tens...

Many-Body Approximation for Tensors

We propose a nonnegative tensor decomposition with focusing on the relat...

Partially Observed Dynamic Tensor Response Regression

In modern data science, dynamic tensor data is prevailing in numerous ap...

Boosted Sparse and Low-Rank Tensor Regression

We propose a sparse and low-rank tensor regression model to relate a uni...

Tensor-on-tensor regression

We propose a framework for the linear prediction of a multi-way array (i...

Sparse Symmetric Tensor Regression for Functional Connectivity Analysis

Tensor regression models, such as CP regression and Tucker regression, h...