[1]Department of Food Science, University of Copenhagen, Denmark.[2]Eigenvector Research, Inc., Manson, WA, USA.
The SIT (Shift-Invariant Tri-linear Tensor model) algorithm decomposes complex tensor datasets, particularly from gas chromatography-mass spectrometry (GC-MS). It offers a 20-60x speed advantage over the latest PARAFAC2 implementations, allows constraints across all modes during optimization, and requires fewer latent variables than existing methods while achieving robust real-world performance.
