Hello,
in my data set I have species densities (continuous data) and some percentage cover (estimated, categorical). I would like to run a multivariate analysis and look at species composition across sites, but is there a way to combine these two type of variables? Can we analyse numbers and categories at the same time in a multivariate analysis? Thank you for your help
Typicall yes. ANOVAS are appropriate for mixed models, as well as dummy-coding in OLS regression.
SEM and IRT and also appropriate for categor-continuous variables analyses within-model.
Hope this helps,
Matt
Hi,
Yes it is possible to analyse categorical and continuous variables together. For more information please go through this link:
https://www.researchgate.net/publication/267547523_Pure_Categorical_Optimization__a_Global_Descent_Approach
You can compute Gower distances between your samples and then compute a PCOA or nMDS. This distance can handle continuous, ordinal, nominal or binary variables and even tolerates NAs
Mirta Zupan try Boosted Regression Trees - machine learning algorithm.
It will fit nicely your design! Good luck.
You can look at Hill-Smith ordinations : https://pbil.univ-lyon1.fr/ADE-4/ade4-html/dudi.hillsmith.html
- Badges
- Science topic