Multi-way data analysis methods like PARAFAC (CP) and PARAFAC2 are commonly used for decomposing complex higher-order tensor datasets, such as those generated by gas chromatography coupled with mass spectrometric detection (GC-MS). These methods are particularly useful for handling retention time shifts and shape changes in chromatographic elution profiles. However, they come with limitations, such as slow computational speed and difficulty in applying constraints like non-negativity to certain modes.
The Shift-Invariant Tri-linearity (SIT) algorithm addresses these issues. It offers a significant speed advantage, being 20–60 times faster than the latest implementations of PARAFAC2. Additionally, SIT allows for the application of constraints across all modes, offering greater flexibility in model optimization. While it doesn’t model shape changes in elution profiles, its robustness in real-world applications has been demonstrated to be high. SIT also tends to require fewer latent variables to extract the same chemical information, sometimes achieving better factor resolution than existing benchmarks.
How to cite
Schneide, P-A, Bro, R, Gallagher, NB. Shift-invariant tri-linearity—A new model for resolving untargeted gas chromatography coupled mass spectrometry data. Journal of Chemometrics. 2023; 37(8):e3501. doi:10.1002/cem.3501
