I often need to compare two or more matrix samples to check if they are similar along a composition profile. Each component is unique and independent from the others. It is common practice to expressed each component mass as % of the total mass. So variables are ratios, continuous and positive, and they sums to 1. Which test can be applied to check if they can be considered as taken from the same substance?
An example of the data is attached.
Pl read the relevant paper
http://www.pelagicos.net/MARS6300/readings/Clarke_et_al._2008.pdf
Thanks very much Ajit kumar Roy, this paper promises to be the solution of my problem, it is much more relevant than all I was able to find up to now... I will follow it carefully to verify if it applies successfully on my data.
The paper bases the comparison between samples on similarity (Bray-Curtis) with unknown probability distribution. The distribution is given by a permutation method with advantages and disadvantages. I would like know if does it exist (or not) a parametric or probabilistic approach. Does it?
- Badges
- Science topic
- Similar topics
- Mathematics
- Algebra
- Matrix