I am trying to evaluate the sensitivity of some of the most commonly used Biodiversity Indices (for details see http://dx.doi.org/10.1002/iroh.200390004 [Beisel et al. 2003]) and I want to create random subsets from my orginal dataset.
I am currently using MS Excel as a basis for storing and calculating data, then switching to Statgraphics Centurion 16 for statistical analyses and PRIMER 5.1.2 for the Biodiversity estimations.
I was wondering if there's a specific software that does such simulations or if anyone has done something similar with a series of commands through another software (e.g. EXCEL formulas).
Thank you all in advance!
Dear Alexis,
I do not know about especific softwares, mostly because I am not from your research area. But I use R software to a number of simulation problems. And is not so hard to learn!
You need the parameters of your actual population and an ideia of the parameters to be used in your new populations. Most important, what is the general aspects of your population distribution?
Knowing this, you can use functions like rnorm, runif, rbeta, etc to simulate how many populations you want easily! All other statistics can be computed with the same algorithm. If you don't have experience with algorithm, don't be intimidated! Use help(rnorm) to start!
Thanks Sandro!
R really crossed my mind, but as you said, I feel rather intimated! I tried lots of times to start learning, especially to create graphs in vector format..but I always got overwhelmed!! If I don't find another alternative, I might give it another try...but then I warn you....I might overwhelm you with questions!!!!
:)
Best regards!
For instance, just to start, if you want to simulate a population of 1000 individuals with Normal distribution with mean equal to 10 and standard deviation of 5, you can simple write:
x = rnorm(1000, 10, 5)
Of course, more complicated distributions will need harder work... ;) Try the help as I said before and click on "distributions" link
The problem is that I don't want to "create" new populations... I already have a dataset and just want to randomly create subsets from the original. For example, my original dataset has 6 different species of marine invertebrates with 2, 3, 6, 7, 3, 5 individuals each... I want to create random datasets, for example a dataset with 3 species and 2, 6, 7 individuals, another one with 4 species with 2, 6, 3, 5 individuals and so on.....
Imagine now having 80 species with different number of individuals and try to create 50 subsets with a random number of these species with their respective populations...
:)
It seems to become easier! What you want is just to resample from your original population. You need to defina a number "N" of neww populations. To each population, you need to define a number "S" of species (let's say a random number) and a vector of size S with the number "I" of individuals per species. Now, you use the "sample" function to randomly select individuals from your original population (with or without replacement, is up to you).
Put it all together in a loop and it is ready!!! :D
Hi Alexis,
Have you looked at Pop-tools for excel (http://www.poptools.org)? It is an add-in for excel with some really intuitive resampling statistics. It is not a comprehensive solution like R (which is probably the way to go if you have some time), but may help you out.
Thanks Pedro! I will definitely give it a try since it's free to use! I think I found a solution to my problem which involved some coding in VBA in EXCEL.
Thank you both for your time and your suggestions!!!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics