CorrIndex: A permutation invariant performance index

Permutation and scaling ambiguities are relevant issues in tensor decomposition source separation algorithms. Although these inevitable when working on real data sets, it is preferred to eliminate uncertainties for evaluating algorithms synthetic sets. As shown the paper, existing performance indices this purpose either greedy unreliable or computationally costly. In we propose a new index, called CorrIndex, whose reliability can be proved theoretically. Moreover, compared previous indices, has low computational cost. Theoretical results computer experiments demonstrate advantages of CorrIndex other indices.