Multi-way PCAA customer recently asked if JMP can do a three-way PCA like what is discussed in the attached article. The authors used a Tucker3 algorithm in Matlab to do PCA on a multi-way data arrays. I didn't find anything in JMP documentation so I think the answer is, no, JMP can't do this. I am hoping to confirm this conclusion and see if anyone has any insight about this type of analysis.
Thanks! -Andrea
Re: Multi-way PCA@andreacoombs1, This looks like an interesting technique. Do you know the intended application for MPCA? The applications I have seen in the literature it seems like FDE might be a good fit at least for data processing. https://www.sciencedirect.com/science/article/pii/S1474667015382719#bbib7 This paper for instance requires a lot of data processing to be able to use the technique. It looks like generating functional principal components than using the model-driven multivariate control chart would be much quicker and easier.
Re: Multi-way PCA@phersh They are trying to link process data from coffee roasting (one of my favorite subjects) with sensory data. They collect a lot of data over time in their process so we are going to be talking about FDE in January. They are particularly interested in understanding the production of aromatic compounds over time and linking it with sensory data. Multiway PCA came up during our initial conversation of FDE.
Your comments about using FDE for data processing then using Multivariate Control Charts is very interesting. I may have to give you a call and discuss further.
Re: Multi-way PCAFeeling some jealousy Andrea! Sounds like a really fun problem. Let us know what they end up doing:)
