Let's say I have data of body size for mammal species. I can classify this data according to biome's name, but not according to the geographic location for each body size. Now, I want to know how body size is distributed within biomes. How do I use ripley's k or nearest neighbor function in this case?
This article, from the Encyclopedia of Environmetrics, talks about differences between Ripley's K and nearest neighbor:
http://www.public.iastate.edu/~pdixon/stat406/RipleyK.pdf
If I understand your question, you just have a set of body size measurements with a given biome. How many numbers constitute each measurement? If the dimensionality is low, you might also just try looking at multivariate mean, covariance, density estimate, etc.. Whatever the dimensionality (including dimension 1), you can measure Euclidian distance (or some other distance function) between data points, and treat it like spatial data. Thus Ripley's K and similar techniques can be used. I wouldn't think you'd expect to find complex "spatial" patterns in body size data, though. It also depends on what your research questions are.
Thanks David, yes you are right I have a set of body size within a given biogeography. I was thinking of finding distribution pattern of body size, if it is bimodal, multimodal or monotonic. I tried with the kernel density but I think kernel density does not include spatial relation. So, looking for one dimension ripley's k function if that works or nearest neighbour density estimation function.
- Badges
- Science topic
- Similar topics
- Medicine
- Physiology
- Biosignals
- Biomedical Signal Processing
- Statistical Analysis