TLViz: Visualising and analysing tensor decomposition models with Python
Peer reviewed, Journal article
Published version
![Thumbnail](/oda-xmlui/bitstream/handle/11250/3061444/10.21105.joss.04754.pdf.jpg?sequence=6&isAllowed=y)
View/ Open
Date
2022Metadata
Show full item recordCollections
Original version
https://doi.org/10.21105/joss.04754Abstract
Multi-way data, also known as tensor data or data cubes, occur in many applications, such as text mining (Bader et al., 2008), neuroscience (Andersen & Rayens, 2004) and chemical analysis (Bro, 1997). Uncovering the meaningful patterns within such data can provide crucial insights into the data source, and tensor decompositions have proven an effective tool for this task. In particular, the PARAFAC model, also known as CANDECOMP/PARAFAC (CP) or the canonical polyadic decomposition (CPD), has shown great promise for extracting interpretable components. PARAFAC has, for example, extracted topics from an email corpus (Bader et al., 2008) and chemical spectra from fluorescence spectroscopy data (Bro, 1997). For a thorough introduction to tensor methods, we refer the reader to (Tamara G. Kolda & Bader, 2009) and (Bro, 1997). The goal of TensorLy-Visualisation (TLViz) is to provide utilities for analysing, visualising and working with tensor decompositions for data analysis in Python.