Im comparing the results of some genetic programming, and I'd like to compare results of 30 runs with 7 sifferent settings (totaling 210 samples).
the distribution is non uniform
Can I use kruskall wallis to do so ?
I've implemented it like the atach I placed here, the result is (from an R snipped i created):
data: aj and y
Kruskal-Wallis chi-squared = 14.8564, df = 6, p-value = 0.0214
can I say the groups are different ?
You would use the Kruskal Wallis test to compare two or more samples. I think in your case should be OK to use this test. The Kruskal Wallis test does not make assumptions about normality. However, it assumes that the observations in each group come from populations with the same shape of distribution. In KNIME Analytics Platform the test is applied at a significance level of 5%. Therefore, you will find statistical difference between groups if p-value is < 0.05. So, in your case this means that the groups are different.
Hope that helps,