Modelling hidden structure of signals in group data analysis with modified (Lr, 1) and block-term decompositions

TL;DR

This paper introduces a novel tensor decomposition approach using (Lr, 1) and Tucker blocks for modeling group data, demonstrating its effectiveness in classification and clustering tasks compared to existing matrix models.

Contribution

It presents a new generalization of block tensor decomposition tailored for group data analysis, expanding the modeling capabilities beyond traditional matrix methods.

Findings

Improved multilabel classification accuracy with the proposed tensor models.

Enhanced clustering performance over known matrix-based methods.

Demonstrated the applicability of tensor decompositions in group activity modeling.

Abstract

This work is devoted to elaboration on the idea to use block term decomposition for group data analysis and to raise the possibility of modelling group activity with (Lr, 1) and Tucker blocks. A new generalization of block tensor decomposition was considered in application to group data analysis. Suggested approach was evaluated on multilabel classification task for a set of images. This contribution also reports results of investigation on clustering with proposed tensor models in comparison with known matrix models, namely common orthogonal basis extraction and group independent component analysis.