Dear all,
I am working on Modification of meta heuristics. I need to perform two-sided Wilcoxon signed-rank test to check statistical test of CEC 2005.
Here, I have best results, mean results and standard deviations of other meta heuristics.
Please share your view about this test.
Thanks
Thanks Vidhi,
It gives p-Value, but how to get T and W.
Can we do this in Exl.
http://dx.doi.org/10.1016/j.asoc.2015.03.003
I got following type of answers on MATLAB.
p-
h
z
ranksum
But I am not getting significance of each value.
Please add your views
You can check this paper, perhaps it helps. In Matlab I've used the signrank() method, but I only cared about the p-value and ranksum.
Article A practical tutorial on the use of nonprametric statistical ...
Dear Antonio,
Thanks for sharing the paper.
I tried this solution in Matlab using [p,h,stats]=ranksum(x,y) code. But I got different results for p value
i.e.
Table 6 shown SaDE versus PSO p-value as 0.00673.
However, using Matlab I got following result:
p =0.3224
h =0
stats =
zval: -0.9896
ranksum: 586
I am not getting similar result. What is wrong with this.
Please add your solution.
Here, I considered following matlab code:
x=[8.42E-09 8.21E-09 6.56E+03 8.09E-09 8.64E-09 1.61E-02 1.26E+03 2.03E+01 8.33E-09 1.55E+01 6.80E+00 5.63E+01 7.07E+01 3.42E+00 8.42E+01 1.23E+02 1.39E+02 5.32E+02 5.20E+02 4.77E+02 5.14E+02 7.66E+02 6.51E+02 2.00E+02 1.74E+03]; % values for SaDE (Table )
y=[1.23E-04 2.60E-02 5.17E+04 2.49E+00 4.10E+02 7.31E+02 2.68E+01 2.04E+01 1.44E+01 1.40E+01 5.59E+00 6.36E+02 1.50E+00 3.30E+00 3.40E+02 1.33E+02 1.50E+02 8.51E+02 8.50E+02 8.51E+02 9.14E+02 8.07E+02 1.03E+03 4.12E+02 5.10E+02];
[p,h,stats]=ranksum(x,y)
Dear Barbara Sawicka,
Thanks for sharing your kind knowledge.
Your answer is remarkable and useful to me. However I need more explanation to clarify my doubts.
I request please share some examples or related work.
Thanks ones again
Please share your knowledge
Hi Ghanshyam G Tejani,
You can use SPSS to perform the Wilcoxon signed-rank test. You can do it, by follow these steps:
1- Analyze> Nonparametric Tests> 2 related samples
2- in the test type, select Wilcoxon.
I tried the date you provided and the P=0.045.
Best regards.
- Badges
- Science topic
- Similar topics
- Computer Communications (Networks)
- Simulators