Loading...

CorrIndex: A permutation invariant performance index

Sobhani, E ; Sharif University of Technology | 2022

90 Viewed
  1. Type of Document: Article
  2. DOI: 10.1016/j.sigpro.2022.108457
  3. Publisher: Elsevier B.V , 2022
  4. Abstract:
  5. Permutation and scaling ambiguities are relevant issues in tensor decomposition and source separation algorithms. Although these ambiguities are inevitable when working on real data sets, it is preferred to eliminate these uncertainties for evaluating algorithms on synthetic data sets. As shown in the paper, the existing performance indices for this purpose are either greedy and unreliable or computationally costly. In this paper, we propose a new performance index, called CorrIndex, whose reliability can be proved theoretically. Moreover, compared to previous performance indices, it has a low computational cost. Theoretical results and computer experiments demonstrate these advantages of CorrIndex compared to other indices. © 2022 Elsevier B.V
  6. Keywords:
  7. Assignment ; Blind source separation ; Hungarian algorithm ; ICA ; Maximum weighted matching ; Optimal transport ; Performance index ; Permutation and scaling ambiguity ; Tensor decomposition ; Blind source separation ; Tensors ; Assignment ; Hungarian algorithm ; Maximum weighted matching ; Optimal transport ; Performance indices ; Permutation and scaling ambiguity ; Real data sets ; Scalings ; Separation algorithms ; Tensor decomposition ; Optimal systems
  8. Source: Signal Processing ; Volume 195 , 2022 ; 01651684 (ISSN)
  9. URL: https://www.sciencedirect.com/science/article/abs/pii/S0165168422000044