in the last days i was reading a laura pla, a biostatistics expert and she states that shannon index isnt asymptotically normal (DOI: 10.1198/1085711043136)
but why?
please explain to me
The distribution of H is be asymptotically normal, but under a large range of typical conditions it is far from being reasonably close to a normal distribution. Therefor it is typically not advisable to rely on the normal approximation. A nice simulation example is shown be Wolfgang Huber here:
https://stats.stackexchange.com/questions/25235/what-test-should-i-use-with-shannon-diversity-index-values
yea, i get what meants asymptotically, but she states that the H' istn asymptotically normal
She does not give any reference, proof nor evidence, what is not good. It might have to do with her point that not all species are sampled... not sure. Best to ask her directly. Her Email is accessible via the publication.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Control Theory
- Control Systems