Hello all,
I have several plots that I would like to do statistics on. All plots are based on 173 data points, derived from movies of organ cultures that are differently treated. My simple statistical mind first thought to just go ahead and compare every single data point of the curves, but that sounds overly complicated and time consuming, and so I was wondering if there's a better way to do it. From my search online it looks like the Pompeiu-Hausdorff method is best suited for my needs, but before I go and learn a whole new statistical test, I'd like to confirm that I'm right. What would you all suggest? (so to clarify, I'm not comparing formulas, I'm comparing series of data points)
Hi Benjamin,
It is not clear to me what you want to achieve with your analysis. If what you want is to find out how close two time-series are from each other (an important problem in Data Mining) I would suggest the R package "TSdist" which can be found here: https://cran.r-project.org/web/packages/TSdist/index.html.
Hope it helps,
Augusto
Hi Augusto, that does sound like what I might need!
To clarify, I have curves that reflect the increasing intensity of a fluorescent marker, and I would like to test whether these curves are significantly different. Will this R package accomplish what I want to do?
Hi Benjamin,
Yes, this package presents different techniques to calculate the distance between time series, some of those techniques should be all right for what you want. A comprehensive discussion of the topic is presented in the attached paper by Mori et al.
Hope it helps,
Augusto
Thanks Augusto!
The one thing I don't quite understand is how distance measurements will give me a p-value. Does TSdist have a function that will give me a p-value, to answer the questions whether my curves are significantly different, or am I being very naive right now?