Dear all,
I want to compare differences in the drainage capillary pressure curves for different simulations with experimental data. I want to show how close or far I am from the experimental data. What would be the most elegant solution? The orange line is the capillary pressure from experimental data.
Thank you in advance!
Best,
Alex
Alexandru - are these curves probability distributions, interpolation polynomials, or something else? Because the answers you've got up to now come from statisticians, and they assumed that the curves are indeed probability distributions.
I don't really understand how rank correlation could be helpful here. If both curves (experimental and simulated) are strinktly monotonousley increasing, the rank correlation will be perfect, no matter how similar the two curves are.
I think the Kolmogorov-Sminov statistic should be a natural choice (https://en.wikipedia.org/wiki/Kolmogorov%E2%80%93Smirnov_test).
If you can take the saturation (as a variable strincly in (0,1)) as a probability value, you might even use the Kulback-Leibler divergence to measure the (dis-)similarity between two curves (https://en.wikipedia.org/wiki/Kullback%E2%80%93Leibler_divergence).
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics